Instructions to use gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection with Transformers:

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

pipe = pipeline("text-generation", model="gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection")
model = AutoModelForMultimodalLM.from_pretrained("gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection

SGLang

How to use gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection 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 "gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection" \
    --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": "gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection" \
        --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": "gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection with Docker Model Runner:
```
docker model run hf.co/gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection
```

zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the FFN-KV Injection (Train-free FFN gating) merge method using meta-llama/Meta-Llama-3-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

shenzhi-wang/Llama3-8B-Chinese-Chat

Configuration

The following YAML configuration was used to produce this model:

dtype: bfloat16
tokenizer:
  source: union
merge_method: ffn_kv_injection
base_model: meta-llama/Meta-Llama-3-8B-Instruct
models:
- model: meta-llama/Meta-Llama-3-8B-Instruct
  parameters: {}
- model: shenzhi-wang/Llama3-8B-Chinese-Chat
  parameters: {}
parameters:
  weights:
  - 0.7
  - 0.3
  tau: 1.0
  aspect_thresh: 1.5
  pnorm: 2.0
  alpha_floor: 0.0
  alpha_ceil: 1.0
write_readme: README.md

Downloads last month: 1

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for gsjang/zh-llama3-8b-chinese-chat-x-meta-llama-3-8b-instruct-ffn_kv_injection

meta-llama/Meta-Llama-3-8B-Instruct

shenzhi-wang/Llama3-8B-Chinese-Chat

Merge model

this model