Instructions to use rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm")
model = AutoModelForMultimodalLM.from_pretrained("rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm

SGLang

How to use rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm with Docker Model Runner:
```
docker model run hf.co/rdtand/Qwen3.6-27B-PrismaQuant-5.5bit-vllm
```

rdtand commited on Apr 22

Commit

09de726

verified ·

1 Parent(s): e7c8b12

v5: max-not-sum sibling aggregation, kernel shape mask, joint input_global — validator: ppl=4.16, mean_NLL=1.43, MTP P0=89.5%

Browse files

Files changed (8) hide show

config.json +76 -76
mixed_native_manifest.json +3 -3
model-00002-of-00006.safetensors +2 -2
model-00003-of-00006.safetensors +2 -2
model-00004-of-00006.safetensors +2 -2
model-00005-of-00006.safetensors +2 -2
model-00006-of-00006.safetensors +2 -2
model.safetensors.index.json +148 -113

config.json CHANGED Viewed

@@ -165,60 +165,44 @@
           "zp_dtype": "torch.uint8"
         },
         "targets": [
-          "re:^language_model[.]model[.]layers[.]11[.]self_attn[.]k_proj$",
-          "re:^language_model[.]model[.]layers[.]11[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]11[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]11[.]self_attn[.]v_proj$",
-          "re:^language_model[.]model[.]layers[.]15[.]self_attn[.]k_proj$",
-          "re:^language_model[.]model[.]layers[.]15[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]15[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]15[.]self_attn[.]v_proj$",
-          "re:^language_model[.]model[.]layers[.]27[.]self_attn[.]o_proj$",
-          "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]in_proj_qkv$",
-          "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]in_proj_qkvz$",
-          "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]29[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]29[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]29[.]linear_attn[.]in_proj_z$",
-          "re:^language_model[.]model[.]layers[.]32[.]mlp[.]gate_proj$",
-          "re:^language_model[.]model[.]layers[.]32[.]mlp[.]gate_up_proj$",
-          "re:^language_model[.]model[.]layers[.]32[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]35[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]44[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]44[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]44[.]linear_attn[.]in_proj_z$",
-          "re:^language_model[.]model[.]layers[.]45[.]mlp[.]gate_proj$",
-          "re:^language_model[.]model[.]layers[.]45[.]mlp[.]gate_up_proj$",
-          "re:^language_model[.]model[.]layers[.]45[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]46[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]47[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]56[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]56[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]56[.]linear_attn[.]in_proj_z$",
-          "re:^language_model[.]model[.]layers[.]59[.]self_attn[.]k_proj$",
-          "re:^language_model[.]model[.]layers[.]59[.]self_attn[.]o_proj$",
-          "re:^language_model[.]model[.]layers[.]59[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]59[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]59[.]self_attn[.]v_proj$",
-          "re:^language_model[.]model[.]layers[.]63[.]mlp[.]down_proj$",
-          "re:^language_model[.]model[.]layers[.]7[.]self_attn[.]k_proj$",
-          "re:^language_model[.]model[.]layers[.]7[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]7[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]7[.]self_attn[.]v_proj$",
-          "re:^visual[.]blocks[.]10[.]attn[.]qkv$",
-          "re:^visual[.]blocks[.]11[.]attn[.]proj$",
-          "re:^visual[.]blocks[.]11[.]attn[.]qkv$",
-          "re:^visual[.]blocks[.]12[.]attn[.]qkv$",
-          "re:^visual[.]blocks[.]13[.]attn[.]proj$",
-          "re:^visual[.]blocks[.]14[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]6[.]attn[.]qkv$",
-          "re:^visual[.]blocks[.]7[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]9[.]attn[.]qkv$"
         ]
       },
@@ -324,6 +308,9 @@
           "re:^language_model[.]model[.]layers[.]17[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]17[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]17[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_z$",
@@ -336,11 +323,7 @@
           "re:^language_model[.]model[.]layers[.]19[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]19[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]19[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]19[.]self_attn[.]k_proj$",
           "re:^language_model[.]model[.]layers[.]19[.]self_attn[.]o_proj$",
-          "re:^language_model[.]model[.]layers[.]19[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]19[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]19[.]self_attn[.]v_proj$",
           "re:^language_model[.]model[.]layers[.]2[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]2[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]2[.]linear_attn[.]in_proj_z$",
@@ -349,6 +332,9 @@
           "re:^language_model[.]model[.]layers[.]2[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]2[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]2[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_z$",
@@ -383,11 +369,7 @@
           "re:^language_model[.]model[.]layers[.]23[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]23[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]23[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]k_proj$",
           "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]o_proj$",
-          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]v_proj$",
           "re:^language_model[.]model[.]layers[.]24[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]24[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]24[.]linear_attn[.]in_proj_z$",
@@ -416,6 +398,10 @@
           "re:^language_model[.]model[.]layers[.]27[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]27[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]27[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]28[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]28[.]mlp[.]gate_proj$",
@@ -442,6 +428,9 @@
           "re:^language_model[.]model[.]layers[.]31[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]32[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]32[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]33[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]33[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]33[.]mlp[.]gate_proj$",
@@ -456,6 +445,10 @@
           "re:^language_model[.]model[.]layers[.]35[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]35[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]35[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_z$",
@@ -487,11 +480,7 @@
           "re:^language_model[.]model[.]layers[.]39[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]39[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]39[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]k_proj$",
           "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]o_proj$",
-          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]v_proj$",
           "re:^language_model[.]model[.]layers[.]4[.]linear_attn[.]in_proj_a$",
           "re:^language_model[.]model[.]layers[.]4[.]linear_attn[.]in_proj_b$",
           "re:^language_model[.]model[.]layers[.]4[.]linear_attn[.]in_proj_ba$",
@@ -538,8 +527,13 @@
           "re:^language_model[.]model[.]layers[.]44[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]45[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]45[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]48[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]48[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]48[.]mlp[.]gate_proj$",
@@ -570,11 +564,7 @@
           "re:^language_model[.]model[.]layers[.]51[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]mlp[.]up_proj$",
-          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]k_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]o_proj$",
-          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]q_proj$",
-          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]qkv_proj$",
-          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]v_proj$",
           "re:^language_model[.]model[.]layers[.]52[.]linear_attn[.]in_proj_a$",
           "re:^language_model[.]model[.]layers[.]52[.]linear_attn[.]in_proj_b$",
           "re:^language_model[.]model[.]layers[.]52[.]linear_attn[.]in_proj_ba$",
@@ -642,6 +632,7 @@
           "re:^language_model[.]model[.]layers[.]59[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]59[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]59[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]6[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]6[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]6[.]linear_attn[.]in_proj_z$",
@@ -680,6 +671,7 @@
           "re:^language_model[.]model[.]layers[.]62[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]62[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]62[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]63[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]7[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]7[.]mlp[.]gate_proj$",
@@ -717,16 +709,23 @@
           "re:^mtp[.]layers[.]0[.]self_attn[.]q_proj$",
           "re:^mtp[.]layers[.]0[.]self_attn[.]qkv_proj$",
           "re:^mtp[.]layers[.]0[.]self_attn[.]v_proj$",
           "re:^visual[.]blocks[.]10[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]10[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]11[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]11[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]12[.]attn[.]proj$",
           "re:^visual[.]blocks[.]12[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]12[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]13[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]13[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]14[.]attn[.]proj$",
           "re:^visual[.]blocks[.]14[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]14[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]15[.]attn[.]proj$",
@@ -766,15 +765,18 @@
           "re:^visual[.]blocks[.]24[.]attn[.]proj$",
           "re:^visual[.]blocks[.]24[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]24[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]25[.]attn[.]proj$",
           "re:^visual[.]blocks[.]25[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]25[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]26[.]attn[.]proj$",
           "re:^visual[.]blocks[.]26[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]26[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]5[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]6[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]6[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]7[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]7[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]8[.]mlp[.]linear_fc1$",
@@ -796,6 +798,10 @@
       "language_model.model.layers.10.linear_attn.in_proj_a",
       "language_model.model.layers.10.linear_attn.in_proj_b",
       "language_model.model.layers.10.linear_attn.in_proj_ba",
       "language_model.model.layers.12.linear_attn.in_proj_a",
       "language_model.model.layers.12.linear_attn.in_proj_b",
       "language_model.model.layers.12.linear_attn.in_proj_ba",
@@ -805,21 +811,23 @@
       "language_model.model.layers.14.linear_attn.in_proj_a",
       "language_model.model.layers.14.linear_attn.in_proj_b",
       "language_model.model.layers.14.linear_attn.in_proj_ba",
       "language_model.model.layers.16.linear_attn.in_proj_a",
       "language_model.model.layers.16.linear_attn.in_proj_b",
       "language_model.model.layers.16.linear_attn.in_proj_ba",
       "language_model.model.layers.17.linear_attn.in_proj_a",
       "language_model.model.layers.17.linear_attn.in_proj_b",
       "language_model.model.layers.17.linear_attn.in_proj_ba",
-      "language_model.model.layers.18.linear_attn.in_proj_a",
-      "language_model.model.layers.18.linear_attn.in_proj_b",
-      "language_model.model.layers.18.linear_attn.in_proj_ba",
       "language_model.model.layers.2.linear_attn.in_proj_a",
       "language_model.model.layers.2.linear_attn.in_proj_b",
       "language_model.model.layers.2.linear_attn.in_proj_ba",
-      "language_model.model.layers.20.linear_attn.in_proj_a",
-      "language_model.model.layers.20.linear_attn.in_proj_b",
-      "language_model.model.layers.20.linear_attn.in_proj_ba",
       "language_model.model.layers.24.linear_attn.in_proj_a",
       "language_model.model.layers.24.linear_attn.in_proj_b",
       "language_model.model.layers.24.linear_attn.in_proj_ba",
@@ -850,7 +858,6 @@
       "language_model.model.layers.30.linear_attn.in_proj_qkvz",
       "language_model.model.layers.30.linear_attn.in_proj_z",
       "language_model.model.layers.31.self_attn.k_proj",
-      "language_model.model.layers.31.self_attn.o_proj",
       "language_model.model.layers.31.self_attn.q_proj",
       "language_model.model.layers.31.self_attn.qkv_proj",
       "language_model.model.layers.31.self_attn.v_proj",
@@ -876,9 +883,6 @@
       "language_model.model.layers.35.self_attn.q_proj",
       "language_model.model.layers.35.self_attn.qkv_proj",
       "language_model.model.layers.35.self_attn.v_proj",
-      "language_model.model.layers.36.linear_attn.in_proj_a",
-      "language_model.model.layers.36.linear_attn.in_proj_b",
-      "language_model.model.layers.36.linear_attn.in_proj_ba",
       "language_model.model.layers.38.linear_attn.in_proj_a",
       "language_model.model.layers.38.linear_attn.in_proj_b",
       "language_model.model.layers.38.linear_attn.in_proj_ba",
@@ -892,7 +896,6 @@
       "language_model.model.layers.42.linear_attn.in_proj_qkvz",
       "language_model.model.layers.42.linear_attn.in_proj_z",
       "language_model.model.layers.43.self_attn.k_proj",
-      "language_model.model.layers.43.self_attn.o_proj",
       "language_model.model.layers.43.self_attn.q_proj",
       "language_model.model.layers.43.self_attn.qkv_proj",
       "language_model.model.layers.43.self_attn.v_proj",
@@ -942,6 +945,10 @@
       "language_model.model.layers.56.linear_attn.in_proj_a",
       "language_model.model.layers.56.linear_attn.in_proj_b",
       "language_model.model.layers.56.linear_attn.in_proj_ba",
       "language_model.model.layers.6.linear_attn.in_proj_a",
       "language_model.model.layers.6.linear_attn.in_proj_b",
       "language_model.model.layers.6.linear_attn.in_proj_ba",
@@ -951,10 +958,10 @@
       "language_model.model.layers.63.mlp.gate_proj",
       "language_model.model.layers.63.mlp.gate_up_proj",
       "language_model.model.layers.63.mlp.up_proj",
-      "language_model.model.layers.63.self_attn.k_proj",
-      "language_model.model.layers.63.self_attn.q_proj",
-      "language_model.model.layers.63.self_attn.qkv_proj",
-      "language_model.model.layers.63.self_attn.v_proj",
       "mtp.fc",
       "visual.blocks.0.attn.proj",
       "visual.blocks.0.attn.qkv",
@@ -964,15 +971,12 @@
       "visual.blocks.1.attn.qkv",
       "visual.blocks.1.mlp.linear_fc1",
       "visual.blocks.1.mlp.linear_fc2",
-      "visual.blocks.10.attn.proj",
-      "visual.blocks.13.attn.qkv",
       "visual.blocks.2.attn.proj",
       "visual.blocks.2.attn.qkv",
       "visual.blocks.2.mlp.linear_fc1",
       "visual.blocks.2.mlp.linear_fc2",
       "visual.blocks.21.mlp.linear_fc2",
       "visual.blocks.23.mlp.linear_fc2",
-      "visual.blocks.24.mlp.linear_fc2",
       "visual.blocks.25.mlp.linear_fc2",
       "visual.blocks.26.mlp.linear_fc2",
       "visual.blocks.3.attn.proj",
@@ -985,12 +989,8 @@
       "visual.blocks.4.mlp.linear_fc2",
       "visual.blocks.5.attn.proj",
       "visual.blocks.5.attn.qkv",
-      "visual.blocks.5.mlp.linear_fc1",
       "visual.blocks.6.attn.proj",
-      "visual.blocks.7.attn.proj",
       "visual.blocks.8.attn.proj",
-      "visual.blocks.8.attn.qkv",
-      "visual.blocks.9.attn.proj",
       "visual.pos_embed"
     ],
     "quantization_status": "compressed"

           "zp_dtype": "torch.uint8"
         },
         "targets": [
+          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]k_proj$",
+          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]q_proj$",
+          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]qkv_proj$",
+          "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]v_proj$",
           "re:^language_model[.]model[.]layers[.]29[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]29[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]29[.]linear_attn[.]in_proj_z$",
+          "re:^language_model[.]model[.]layers[.]31[.]self_attn[.]o_proj$",
+          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]k_proj$",
+          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]q_proj$",
+          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]qkv_proj$",
+          "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]v_proj$",
+          "re:^language_model[.]model[.]layers[.]43[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]44[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]44[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]44[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]k_proj$",
+          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]q_proj$",
+          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]qkv_proj$",
+          "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]v_proj$",
           "re:^language_model[.]model[.]layers[.]56[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]56[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]56[.]linear_attn[.]in_proj_z$",
+          "re:^language_model[.]model[.]layers[.]63[.]self_attn[.]k_proj$",
+          "re:^language_model[.]model[.]layers[.]63[.]self_attn[.]q_proj$",
+          "re:^language_model[.]model[.]layers[.]63[.]self_attn[.]qkv_proj$",
+          "re:^language_model[.]model[.]layers[.]63[.]self_attn[.]v_proj$",
+          "re:^visual[.]blocks[.]13[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]6[.]attn[.]qkv$",
+          "re:^visual[.]blocks[.]7[.]attn[.]proj$",
+          "re:^visual[.]blocks[.]8[.]attn[.]qkv$",
+          "re:^visual[.]blocks[.]9[.]attn[.]proj$",
           "re:^visual[.]blocks[.]9[.]attn[.]qkv$"
         ]
       },
           "re:^language_model[.]model[.]layers[.]17[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]17[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]17[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_a$",
+          "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_b$",
+          "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_ba$",
           "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]18[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]19[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]19[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]19[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]19[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]2[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]2[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]2[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]2[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]2[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]2[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_a$",
+          "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_b$",
+          "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_ba$",
           "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]20[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]23[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]23[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]23[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]23[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]24[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]24[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]24[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]27[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]27[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]27[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]27[.]self_attn[.]o_proj$",
+          "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]in_proj_qkv$",
+          "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]in_proj_qkvz$",
+          "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]28[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]28[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]28[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]31[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]32[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]32[.]mlp[.]down_proj$",
+          "re:^language_model[.]model[.]layers[.]32[.]mlp[.]gate_proj$",
+          "re:^language_model[.]model[.]layers[.]32[.]mlp[.]gate_up_proj$",
+          "re:^language_model[.]model[.]layers[.]32[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]33[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]33[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]33[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]35[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]35[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]35[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]35[.]self_attn[.]o_proj$",
+          "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_a$",
+          "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_b$",
+          "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_ba$",
           "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]36[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]39[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]39[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]39[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]39[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]4[.]linear_attn[.]in_proj_a$",
           "re:^language_model[.]model[.]layers[.]4[.]linear_attn[.]in_proj_b$",
           "re:^language_model[.]model[.]layers[.]4[.]linear_attn[.]in_proj_ba$",
           "re:^language_model[.]model[.]layers[.]44[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]45[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]45[.]mlp[.]down_proj$",
+          "re:^language_model[.]model[.]layers[.]45[.]mlp[.]gate_proj$",
+          "re:^language_model[.]model[.]layers[.]45[.]mlp[.]gate_up_proj$",
+          "re:^language_model[.]model[.]layers[.]45[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]46[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]46[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]47[.]mlp[.]down_proj$",
+          "re:^language_model[.]model[.]layers[.]47[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]48[.]linear_attn[.]out_proj$",
           "re:^language_model[.]model[.]layers[.]48[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]48[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]mlp[.]up_proj$",
           "re:^language_model[.]model[.]layers[.]51[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]52[.]linear_attn[.]in_proj_a$",
           "re:^language_model[.]model[.]layers[.]52[.]linear_attn[.]in_proj_b$",
           "re:^language_model[.]model[.]layers[.]52[.]linear_attn[.]in_proj_ba$",
           "re:^language_model[.]model[.]layers[.]59[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]59[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]59[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]59[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]6[.]linear_attn[.]in_proj_qkv$",
           "re:^language_model[.]model[.]layers[.]6[.]linear_attn[.]in_proj_qkvz$",
           "re:^language_model[.]model[.]layers[.]6[.]linear_attn[.]in_proj_z$",
           "re:^language_model[.]model[.]layers[.]62[.]mlp[.]gate_proj$",
           "re:^language_model[.]model[.]layers[.]62[.]mlp[.]gate_up_proj$",
           "re:^language_model[.]model[.]layers[.]62[.]mlp[.]up_proj$",
+          "re:^language_model[.]model[.]layers[.]63[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]63[.]self_attn[.]o_proj$",
           "re:^language_model[.]model[.]layers[.]7[.]mlp[.]down_proj$",
           "re:^language_model[.]model[.]layers[.]7[.]mlp[.]gate_proj$",
           "re:^mtp[.]layers[.]0[.]self_attn[.]q_proj$",
           "re:^mtp[.]layers[.]0[.]self_attn[.]qkv_proj$",
           "re:^mtp[.]layers[.]0[.]self_attn[.]v_proj$",
+          "re:^visual[.]blocks[.]10[.]attn[.]proj$",
+          "re:^visual[.]blocks[.]10[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]10[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]10[.]mlp[.]linear_fc2$",
+          "re:^visual[.]blocks[.]11[.]attn[.]proj$",
+          "re:^visual[.]blocks[.]11[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]11[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]11[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]12[.]attn[.]proj$",
+          "re:^visual[.]blocks[.]12[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]12[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]12[.]mlp[.]linear_fc2$",
+          "re:^visual[.]blocks[.]13[.]attn[.]proj$",
           "re:^visual[.]blocks[.]13[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]13[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]14[.]attn[.]proj$",
+          "re:^visual[.]blocks[.]14[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]14[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]14[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]15[.]attn[.]proj$",
           "re:^visual[.]blocks[.]24[.]attn[.]proj$",
           "re:^visual[.]blocks[.]24[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]24[.]mlp[.]linear_fc1$",
+          "re:^visual[.]blocks[.]24[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]25[.]attn[.]proj$",
           "re:^visual[.]blocks[.]25[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]25[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]26[.]attn[.]proj$",
           "re:^visual[.]blocks[.]26[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]26[.]mlp[.]linear_fc1$",
+          "re:^visual[.]blocks[.]5[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]5[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]6[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]6[.]mlp[.]linear_fc2$",
+          "re:^visual[.]blocks[.]7[.]attn[.]qkv$",
           "re:^visual[.]blocks[.]7[.]mlp[.]linear_fc1$",
           "re:^visual[.]blocks[.]7[.]mlp[.]linear_fc2$",
           "re:^visual[.]blocks[.]8[.]mlp[.]linear_fc1$",
       "language_model.model.layers.10.linear_attn.in_proj_a",
       "language_model.model.layers.10.linear_attn.in_proj_b",
       "language_model.model.layers.10.linear_attn.in_proj_ba",
+      "language_model.model.layers.11.self_attn.k_proj",
+      "language_model.model.layers.11.self_attn.q_proj",
+      "language_model.model.layers.11.self_attn.qkv_proj",
+      "language_model.model.layers.11.self_attn.v_proj",
       "language_model.model.layers.12.linear_attn.in_proj_a",
       "language_model.model.layers.12.linear_attn.in_proj_b",
       "language_model.model.layers.12.linear_attn.in_proj_ba",
       "language_model.model.layers.14.linear_attn.in_proj_a",
       "language_model.model.layers.14.linear_attn.in_proj_b",
       "language_model.model.layers.14.linear_attn.in_proj_ba",
+      "language_model.model.layers.15.self_attn.k_proj",
+      "language_model.model.layers.15.self_attn.q_proj",
+      "language_model.model.layers.15.self_attn.qkv_proj",
+      "language_model.model.layers.15.self_attn.v_proj",
       "language_model.model.layers.16.linear_attn.in_proj_a",
       "language_model.model.layers.16.linear_attn.in_proj_b",
       "language_model.model.layers.16.linear_attn.in_proj_ba",
       "language_model.model.layers.17.linear_attn.in_proj_a",
       "language_model.model.layers.17.linear_attn.in_proj_b",
       "language_model.model.layers.17.linear_attn.in_proj_ba",
+      "language_model.model.layers.19.self_attn.k_proj",
+      "language_model.model.layers.19.self_attn.q_proj",
+      "language_model.model.layers.19.self_attn.qkv_proj",
+      "language_model.model.layers.19.self_attn.v_proj",
       "language_model.model.layers.2.linear_attn.in_proj_a",
       "language_model.model.layers.2.linear_attn.in_proj_b",
       "language_model.model.layers.2.linear_attn.in_proj_ba",
       "language_model.model.layers.24.linear_attn.in_proj_a",
       "language_model.model.layers.24.linear_attn.in_proj_b",
       "language_model.model.layers.24.linear_attn.in_proj_ba",
       "language_model.model.layers.30.linear_attn.in_proj_qkvz",
       "language_model.model.layers.30.linear_attn.in_proj_z",
       "language_model.model.layers.31.self_attn.k_proj",
       "language_model.model.layers.31.self_attn.q_proj",
       "language_model.model.layers.31.self_attn.qkv_proj",
       "language_model.model.layers.31.self_attn.v_proj",
       "language_model.model.layers.35.self_attn.q_proj",
       "language_model.model.layers.35.self_attn.qkv_proj",
       "language_model.model.layers.35.self_attn.v_proj",
       "language_model.model.layers.38.linear_attn.in_proj_a",
       "language_model.model.layers.38.linear_attn.in_proj_b",
       "language_model.model.layers.38.linear_attn.in_proj_ba",
       "language_model.model.layers.42.linear_attn.in_proj_qkvz",
       "language_model.model.layers.42.linear_attn.in_proj_z",
       "language_model.model.layers.43.self_attn.k_proj",
       "language_model.model.layers.43.self_attn.q_proj",
       "language_model.model.layers.43.self_attn.qkv_proj",
       "language_model.model.layers.43.self_attn.v_proj",
       "language_model.model.layers.56.linear_attn.in_proj_a",
       "language_model.model.layers.56.linear_attn.in_proj_b",
       "language_model.model.layers.56.linear_attn.in_proj_ba",
+      "language_model.model.layers.59.self_attn.k_proj",
+      "language_model.model.layers.59.self_attn.q_proj",
+      "language_model.model.layers.59.self_attn.qkv_proj",
+      "language_model.model.layers.59.self_attn.v_proj",
       "language_model.model.layers.6.linear_attn.in_proj_a",
       "language_model.model.layers.6.linear_attn.in_proj_b",
       "language_model.model.layers.6.linear_attn.in_proj_ba",
       "language_model.model.layers.63.mlp.gate_proj",
       "language_model.model.layers.63.mlp.gate_up_proj",
       "language_model.model.layers.63.mlp.up_proj",
+      "language_model.model.layers.7.self_attn.k_proj",
+      "language_model.model.layers.7.self_attn.q_proj",
+      "language_model.model.layers.7.self_attn.qkv_proj",
+      "language_model.model.layers.7.self_attn.v_proj",
       "mtp.fc",
       "visual.blocks.0.attn.proj",
       "visual.blocks.0.attn.qkv",
       "visual.blocks.1.attn.qkv",
       "visual.blocks.1.mlp.linear_fc1",
       "visual.blocks.1.mlp.linear_fc2",
       "visual.blocks.2.attn.proj",
       "visual.blocks.2.attn.qkv",
       "visual.blocks.2.mlp.linear_fc1",
       "visual.blocks.2.mlp.linear_fc2",
       "visual.blocks.21.mlp.linear_fc2",
       "visual.blocks.23.mlp.linear_fc2",
       "visual.blocks.25.mlp.linear_fc2",
       "visual.blocks.26.mlp.linear_fc2",
       "visual.blocks.3.attn.proj",
       "visual.blocks.4.mlp.linear_fc2",
       "visual.blocks.5.attn.proj",
       "visual.blocks.5.attn.qkv",
       "visual.blocks.6.attn.proj",
       "visual.blocks.8.attn.proj",
       "visual.pos_embed"
     ],
     "quantization_status": "compressed"

mixed_native_manifest.json CHANGED Viewed

@@ -3,10 +3,10 @@
   "source_recipe": "/work/artifacts/layer_config.json",
   "format_histogram": {
     "head_passthrough/BF16": 3,
-    "linear/BF16": 114,
-    "linear/NVFP4": 348,
     "layer_passthrough/BF16": 352,
-    "linear/MXFP8": 34,
     "mtp_linear/NVFP4": 7,
     "mtp_passthrough/BF16": 8
   },

   "source_recipe": "/work/artifacts/layer_config.json",
   "format_histogram": {
     "head_passthrough/BF16": 3,
+    "linear/BF16": 118,
+    "linear/NVFP4": 354,
     "layer_passthrough/BF16": 352,
+    "linear/MXFP8": 24,
     "mtp_linear/NVFP4": 7,
     "mtp_passthrough/BF16": 8
   },

model-00002-of-00006.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f0c825c267e860aae88a95eb8f304c2c8248727bb8e2d200674aca444a3b93fa
-size 4489298320

 version https://git-lfs.github.com/spec/v1
+oid sha256:d5f3d4bca6418ab2d9fc03502124412c85c56b00000d51c2cc98a450b0d4bb13
+size 4516492384

model-00003-of-00006.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:30495eb82723105ddde40df6959eda2c1d754e156ca56dcfbf7624934c18c9b7
-size 4504718112

 version https://git-lfs.github.com/spec/v1
+oid sha256:773d4a9f7bb6583c51a2124c732ad3cfb27430f92409516cf438ce83ec18a7c2
+size 4488233624

model-00004-of-00006.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1c0f2a6b8f68ee6133b5cd504ddbc8447ae79e02a75c17d51c03f774c6b923f
-size 4459780368

 version https://git-lfs.github.com/spec/v1
+oid sha256:1d918713c09bd215109256804998e5b7113ea5126a2822b393ba38f90e4b0d2a
+size 4451868072

model-00005-of-00006.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3b498a777990a1c8efde1a4191a1d5069357542a2be5ab200c1f2bcd44ef5d94
-size 4508367648

 version https://git-lfs.github.com/spec/v1
+oid sha256:ad777d8cef79c3c4c2f3eadcb8feb447a31e2cc839950336ce3fc520cdf1e658
+size 4532286704

model-00006-of-00006.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d1da27c738bd49e1bad742aa3e95cef832031bf88b9672ea8a2ce0ec25738230
-size 2163987616

 version https://git-lfs.github.com/spec/v1
+oid sha256:66dc99205ef90ee732d7e17ba02b4bb0c71116d4df855bb5c0bb066c7dee899e
+size 2137290832

model.safetensors.index.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
   "metadata": {
-    "total_size": 22668632072
   },
   "weight_map": {
     "lm_head.weight": "model-00001-of-00006.safetensors",
@@ -120,16 +120,13 @@
     "model.language_model.layers.11.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.11.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.input_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.weight_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.weight_packed": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.11.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.11.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.12.input_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.12.linear_attn.A_log": "model-00002-of-00006.safetensors",
     "model.language_model.layers.12.linear_attn.conv1d.weight": "model-00002-of-00006.safetensors",
@@ -242,16 +239,13 @@
     "model.language_model.layers.15.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.15.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.input_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.weight_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.weight_packed": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.15.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.15.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.input_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.A_log": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.conv1d.weight": "model-00002-of-00006.safetensors",
@@ -260,23 +254,23 @@
     "model.language_model.layers.16.linear_attn.in_proj_b.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.in_proj_qkv.input_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.in_proj_qkv.weight_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.in_proj_qkv.weight_packed": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.in_proj_qkv.weight_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.in_proj_z.input_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.in_proj_z.weight_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.in_proj_z.weight_packed": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.in_proj_z.weight_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.norm.weight": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.out_proj.input_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.out_proj.weight_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.out_proj.weight_packed": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.linear_attn.out_proj.weight_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.mlp.down_proj.input_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.mlp.down_proj.weight_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.mlp.down_proj.weight_packed": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.mlp.down_proj.weight_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.mlp.gate_proj.input_global_scale": "model-00002-of-00006.safetensors",
-    "model.language_model.layers.16.mlp.gate_proj.weight_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.mlp.gate_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.16.mlp.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.16.mlp.up_proj.input_global_scale": "model-00003-of-00006.safetensors",
@@ -320,8 +314,14 @@
     "model.language_model.layers.18.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.conv1d.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.18.linear_attn.in_proj_a.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.18.linear_attn.in_proj_b.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.in_proj_qkv.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.in_proj_qkv.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.in_proj_qkv.weight_packed": "model-00003-of-00006.safetensors",
@@ -363,23 +363,14 @@
     "model.language_model.layers.19.mlp.up_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.k_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.k_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.k_proj.weight_packed": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.q_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.q_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.q_proj.weight_packed": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.v_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.v_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.v_proj.weight_packed": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.19.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.2.input_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.2.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.2.linear_attn.conv1d.weight": "model-00003-of-00006.safetensors",
@@ -416,8 +407,14 @@
     "model.language_model.layers.20.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.conv1d.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.20.linear_attn.in_proj_a.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.20.linear_attn.in_proj_b.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.in_proj_qkv.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.in_proj_qkv.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.in_proj_qkv.weight_packed": "model-00003-of-00006.safetensors",
@@ -535,22 +532,16 @@
     "model.language_model.layers.23.mlp.up_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.k_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.k_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.k_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.q_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.q_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.q_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.v_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.v_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.23.self_attn.v_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.24.input_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.24.linear_attn.A_log": "model-00003-of-00006.safetensors",
@@ -664,7 +655,9 @@
     "model.language_model.layers.27.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.27.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
@@ -675,9 +668,13 @@
     "model.language_model.layers.28.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_a.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_b.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.28.linear_attn.in_proj_qkv.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_qkv.weight_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.28.linear_attn.in_proj_z.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_z.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.out_proj.input_global_scale": "model-00003-of-00006.safetensors",
@@ -791,6 +788,7 @@
     "model.language_model.layers.31.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
@@ -800,20 +798,24 @@
     "model.language_model.layers.32.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
     "model.language_model.layers.32.linear_attn.in_proj_a.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.32.linear_attn.in_proj_b.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.in_proj_qkv.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.in_proj_z.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.norm.weight": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.out_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.out_proj.weight_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.out_proj.weight_packed": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.linear_attn.out_proj.weight_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.mlp.down_proj.input_global_scale": "model-00003-of-00006.safetensors",
-    "model.language_model.layers.32.mlp.down_proj.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.32.mlp.down_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.down_proj.weight_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.32.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.32.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.up_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.33.input_layernorm.weight": "model-00004-of-00006.safetensors",
@@ -884,7 +886,9 @@
     "model.language_model.layers.35.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.35.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
@@ -893,8 +897,14 @@
     "model.language_model.layers.36.linear_attn.A_log": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.conv1d.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.dt_bias": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.36.linear_attn.in_proj_a.weight": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.36.linear_attn.in_proj_b.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.in_proj_qkv.input_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.in_proj_qkv.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.in_proj_qkv.weight_packed": "model-00004-of-00006.safetensors",
@@ -1006,22 +1016,16 @@
     "model.language_model.layers.39.mlp.up_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.k_proj.input_global_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.k_proj.weight_global_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.k_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.input_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.q_proj.input_global_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.q_proj.weight_global_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.q_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.v_proj.input_global_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.v_proj.weight_global_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.39.self_attn.v_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.4.input_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.4.linear_attn.A_log": "model-00004-of-00006.safetensors",
@@ -1174,6 +1178,7 @@
     "model.language_model.layers.43.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
@@ -1222,9 +1227,13 @@
     "model.language_model.layers.45.mlp.down_proj.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.down_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.down_proj.weight_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.45.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.45.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.up_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.input_layernorm.weight": "model-00004-of-00006.safetensors",
@@ -1236,7 +1245,9 @@
     "model.language_model.layers.46.linear_attn.in_proj_qkv.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.in_proj_z.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.norm.weight": "model-00004-of-00006.safetensors",
-    "model.language_model.layers.46.linear_attn.out_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.out_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.mlp.down_proj.input_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.mlp.down_proj.weight_global_scale": "model-00004-of-00006.safetensors",
@@ -1259,7 +1270,9 @@
     "model.language_model.layers.47.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.47.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
@@ -1395,22 +1408,16 @@
     "model.language_model.layers.51.mlp.up_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.k_proj.input_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.k_proj.weight_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.k_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.input_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.weight_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.q_proj.input_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.q_proj.weight_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.q_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.v_proj.input_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.v_proj.weight_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.51.self_attn.v_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.52.input_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.52.linear_attn.A_log": "model-00005-of-00006.safetensors",
@@ -1666,14 +1673,13 @@
     "model.language_model.layers.59.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.59.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.59.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.59.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.59.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.6.input_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.6.linear_attn.A_log": "model-00005-of-00006.safetensors",
     "model.language_model.layers.6.linear_attn.conv1d.weight": "model-00005-of-00006.safetensors",
@@ -1800,13 +1806,13 @@
     "model.language_model.layers.62.linear_attn.out_proj.input_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.62.linear_attn.out_proj.weight_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.62.linear_attn.out_proj.weight_packed": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.linear_attn.out_proj.weight_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.mlp.down_proj.input_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.mlp.down_proj.weight_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.mlp.down_proj.weight_packed": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.mlp.down_proj.weight_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.mlp.gate_proj.input_global_scale": "model-00005-of-00006.safetensors",
-    "model.language_model.layers.62.mlp.gate_proj.weight_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.62.mlp.gate_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.mlp.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.mlp.up_proj.input_global_scale": "model-00006-of-00006.safetensors",
@@ -1815,20 +1821,25 @@
     "model.language_model.layers.62.mlp.up_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.input_layernorm.weight": "model-00006-of-00006.safetensors",
-    "model.language_model.layers.63.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.mlp.down_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.input_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.mlp.down_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.mlp.down_proj.weight_global_scale": "model-00006-of-00006.safetensors",
@@ -1845,16 +1856,13 @@
     "model.language_model.layers.7.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
-    "model.language_model.layers.7.self_attn.k_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
-    "model.language_model.layers.7.self_attn.q_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
-    "model.language_model.layers.7.self_attn.v_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.8.input_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.8.linear_attn.A_log": "model-00006-of-00006.safetensors",
     "model.language_model.layers.8.linear_attn.conv1d.weight": "model-00006-of-00006.safetensors",
@@ -1957,9 +1965,14 @@
     "model.visual.blocks.1.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.1.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.attn.proj.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.10.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.attn.qkv.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.10.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
@@ -1976,10 +1989,14 @@
     "model.visual.blocks.10.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.proj.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.11.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.qkv.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.11.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
@@ -2001,7 +2018,9 @@
     "model.visual.blocks.12.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.qkv.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.12.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
@@ -2018,10 +2037,13 @@
     "model.visual.blocks.12.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.proj.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.13.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.mlp.linear_fc1.weight_global_scale": "model-00006-of-00006.safetensors",
@@ -2042,7 +2064,9 @@
     "model.visual.blocks.14.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.qkv.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.14.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
@@ -2296,7 +2320,10 @@
     "model.visual.blocks.24.mlp.linear_fc1.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.mlp.linear_fc1.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.mlp.linear_fc2.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.24.mlp.linear_fc2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.norm1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.norm1.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.norm2.bias": "model-00006-of-00006.safetensors",
@@ -2372,7 +2399,10 @@
     "model.visual.blocks.5.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.5.mlp.linear_fc1.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc2.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc2.weight_global_scale": "model-00006-of-00006.safetensors",
@@ -2403,8 +2433,11 @@
     "model.visual.blocks.6.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.proj.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.qkv.bias": "model-00006-of-00006.safetensors",
-    "model.visual.blocks.7.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
@@ -2424,6 +2457,7 @@
     "model.visual.blocks.8.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.mlp.linear_fc1.weight_global_scale": "model-00006-of-00006.safetensors",
@@ -2440,6 +2474,7 @@
     "model.visual.blocks.8.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.proj.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",

 {
   "metadata": {
+    "total_size": 22668647304
   },
   "weight_map": {
     "lm_head.weight": "model-00001-of-00006.safetensors",
     "model.language_model.layers.11.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.input_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.weight_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.weight_packed": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.11.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.12.input_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.12.linear_attn.A_log": "model-00002-of-00006.safetensors",
     "model.language_model.layers.12.linear_attn.conv1d.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.input_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.weight_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.weight_packed": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.15.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.input_layernorm.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.A_log": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.conv1d.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.in_proj_b.weight": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.in_proj_qkv.input_global_scale": "model-00002-of-00006.safetensors",
     "model.language_model.layers.16.linear_attn.in_proj_qkv.weight_global_scale": "model-00002-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_qkv.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_qkv.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_z.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_z.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_z.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_z.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.norm.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.out_proj.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.out_proj.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.out_proj.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.linear_attn.out_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.mlp.down_proj.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.mlp.down_proj.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.mlp.down_proj.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.mlp.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.mlp.gate_proj.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.16.mlp.gate_proj.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.16.mlp.gate_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.16.mlp.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.16.mlp.up_proj.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.conv1d.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_a.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_a.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_a.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_a.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_b.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_b.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_b.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_b.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.in_proj_qkv.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.in_proj_qkv.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.18.linear_attn.in_proj_qkv.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.mlp.up_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.19.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.19.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.19.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.19.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.2.input_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.2.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.2.linear_attn.conv1d.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.conv1d.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_a.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_a.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_a.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_a.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_b.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_b.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_b.weight_packed": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_b.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.in_proj_qkv.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.in_proj_qkv.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.20.linear_attn.in_proj_qkv.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.mlp.up_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.23.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.weight_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.23.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.23.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.23.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.24.input_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.24.linear_attn.A_log": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.27.self_attn.o_proj.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.27.self_attn.o_proj.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.27.self_attn.o_proj.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.27.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_a.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_b.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.28.linear_attn.in_proj_qkv.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.28.linear_attn.in_proj_qkv.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.28.linear_attn.in_proj_qkv.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_qkv.weight_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.28.linear_attn.in_proj_z.input_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.28.linear_attn.in_proj_z.weight_global_scale": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.28.linear_attn.in_proj_z.weight_packed": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.in_proj_z.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.28.linear_attn.out_proj.input_global_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.31.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.31.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.32.linear_attn.dt_bias": "model-00003-of-00006.safetensors",
     "model.language_model.layers.32.linear_attn.in_proj_a.weight": "model-00003-of-00006.safetensors",
     "model.language_model.layers.32.linear_attn.in_proj_b.weight": "model-00003-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.in_proj_qkv.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.in_proj_z.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.norm.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.out_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.out_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.out_proj.weight_packed": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.linear_attn.out_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.down_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.down_proj.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.down_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.gate_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.gate_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.gate_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.up_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.up_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.32.mlp.up_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.mlp.up_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.32.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.33.input_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.35.self_attn.o_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.35.self_attn.o_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.35.self_attn.o_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.35.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.A_log": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.conv1d.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.dt_bias": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_a.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_a.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_a.weight_packed": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_a.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_b.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_b.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_b.weight_packed": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.36.linear_attn.in_proj_b.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.in_proj_qkv.input_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.in_proj_qkv.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.36.linear_attn.in_proj_qkv.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.mlp.up_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.39.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.input_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.39.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.39.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.39.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.4.input_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.4.linear_attn.A_log": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.43.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.43.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.down_proj.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.down_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.45.mlp.gate_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.45.mlp.gate_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.45.mlp.gate_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.45.mlp.up_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.45.mlp.up_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.45.mlp.up_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.mlp.up_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.45.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.input_layernorm.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.in_proj_qkv.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.in_proj_z.weight": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.norm.weight": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.46.linear_attn.out_proj.input_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.46.linear_attn.out_proj.weight_global_scale": "model-00004-of-00006.safetensors",
+    "model.language_model.layers.46.linear_attn.out_proj.weight_packed": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.linear_attn.out_proj.weight_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.mlp.down_proj.input_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.46.mlp.down_proj.weight_global_scale": "model-00004-of-00006.safetensors",
     "model.language_model.layers.47.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.47.self_attn.o_proj.input_global_scale": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.47.self_attn.o_proj.weight_global_scale": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.47.self_attn.o_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.47.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.mlp.up_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.51.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.input_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.weight_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.51.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.51.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.51.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.52.input_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.52.linear_attn.A_log": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.59.self_attn.o_proj.input_global_scale": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.59.self_attn.o_proj.weight_global_scale": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.59.self_attn.o_proj.weight_packed": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.59.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.6.input_layernorm.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.6.linear_attn.A_log": "model-00005-of-00006.safetensors",
     "model.language_model.layers.6.linear_attn.conv1d.weight": "model-00005-of-00006.safetensors",
     "model.language_model.layers.62.linear_attn.out_proj.input_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.62.linear_attn.out_proj.weight_global_scale": "model-00005-of-00006.safetensors",
     "model.language_model.layers.62.linear_attn.out_proj.weight_packed": "model-00005-of-00006.safetensors",
+    "model.language_model.layers.62.linear_attn.out_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.62.mlp.down_proj.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.62.mlp.down_proj.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.62.mlp.down_proj.weight_packed": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.62.mlp.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.62.mlp.gate_proj.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.62.mlp.gate_proj.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.mlp.gate_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.mlp.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.mlp.up_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.mlp.up_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.62.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.63.mlp.down_proj.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.63.mlp.down_proj.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.63.mlp.down_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.mlp.down_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.63.self_attn.k_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.o_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.63.self_attn.q_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.63.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.language_model.layers.63.self_attn.v_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.input_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.mlp.down_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.mlp.down_proj.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.input_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.o_proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.7.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.8.input_layernorm.weight": "model-00006-of-00006.safetensors",
     "model.language_model.layers.8.linear_attn.A_log": "model-00006-of-00006.safetensors",
     "model.language_model.layers.8.linear_attn.conv1d.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.1.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.1.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.attn.proj.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.proj.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.proj.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.attn.qkv.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.qkv.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.qkv.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.10.attn.qkv.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.10.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.proj.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.11.attn.proj.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.11.attn.proj.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.11.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.qkv.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.11.attn.qkv.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.11.attn.qkv.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.11.attn.qkv.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.11.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.qkv.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.12.attn.qkv.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.12.attn.qkv.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.12.attn.qkv.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.12.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.proj.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.13.attn.proj.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.13.attn.proj.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.13.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.attn.qkv.weight": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.13.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.13.mlp.linear_fc1.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.proj.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.qkv.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.14.attn.qkv.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.14.attn.qkv.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.14.attn.qkv.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.14.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.mlp.linear_fc1.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.mlp.linear_fc1.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.mlp.linear_fc2.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.24.mlp.linear_fc2.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.24.mlp.linear_fc2.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.24.mlp.linear_fc2.weight_packed": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.24.mlp.linear_fc2.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.norm1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.norm1.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.24.norm2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc1.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc1.weight_packed": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc1.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc2.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc2.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.5.mlp.linear_fc2.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.6.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.proj.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.proj.weight": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.7.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.qkv.bias": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.7.attn.qkv.input_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.7.attn.qkv.weight_global_scale": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.7.attn.qkv.weight_packed": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.7.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.attn.proj.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.attn.qkv.weight": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.8.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.mlp.linear_fc1.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.mlp.linear_fc1.input_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.mlp.linear_fc1.weight_global_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.8.norm2.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.proj.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.proj.weight": "model-00006-of-00006.safetensors",
+    "model.visual.blocks.9.attn.proj.weight_scale": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.qkv.bias": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.qkv.weight": "model-00006-of-00006.safetensors",
     "model.visual.blocks.9.attn.qkv.weight_scale": "model-00006-of-00006.safetensors",