Instructions to use mkurman/convgpt-v2-b200-full-synth-20h with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mkurman/convgpt-v2-b200-full-synth-20h with Transformers:

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

pipe = pipeline("text-generation", model="mkurman/convgpt-v2-b200-full-synth-20h")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("mkurman/convgpt-v2-b200-full-synth-20h", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mkurman/convgpt-v2-b200-full-synth-20h with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mkurman/convgpt-v2-b200-full-synth-20h"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mkurman/convgpt-v2-b200-full-synth-20h",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/mkurman/convgpt-v2-b200-full-synth-20h

SGLang

How to use mkurman/convgpt-v2-b200-full-synth-20h 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 "mkurman/convgpt-v2-b200-full-synth-20h" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mkurman/convgpt-v2-b200-full-synth-20h",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "mkurman/convgpt-v2-b200-full-synth-20h" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mkurman/convgpt-v2-b200-full-synth-20h",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use mkurman/convgpt-v2-b200-full-synth-20h with Docker Model Runner:
```
docker model run hf.co/mkurman/convgpt-v2-b200-full-synth-20h
```

convgpt-v2-b200-full-synth-20h / checkpoint-134000 /config.json

mkurman

Upload config.json

c215f60 verified 2 months ago

Raw

History Blame Contribute Delete

2.2 kB

	{
	"architectures": [
	"ConvGPTV2ForCausalLM"
	],
	"auto_map": {
	"AutoConfig": "configuration_convgpt_v2.ConvGPTV2Config",
	"AutoModelForCausalLM": "modeling_convgpt_v2.ConvGPTV2ForCausalLM"
	},
	"attention_every": 0,
	"attention_gate_init": -2.0,
	"attention_num_heads": 4,
	"attention_start_layer": 0,
	"bos_token_id": 1,
	"branch_dropout": 0.0,
	"chunk_memory_gate_init": -4.0,
	"chunk_memory_include_current_chunk": false,
	"chunk_memory_size": 64,
	"chunk_memory_token_top_k": 64,
	"chunk_memory_top_k": 2,
	"conv1d_dilations": [
	1,
	2,
	4,
	8
	],
	"conv1d_kernel_size": 3,
	"conv1d_residual_gate_init": -6.0,
	"conv2d_backend": "triton_gather",
	"conv2d_chunk_size": 1024,
	"conv2d_dilations": [
	1,
	2,
	4,
	8
	],
	"conv2d_kernel_size": 3,
	"conv2d_residual_gate_init": 2.0,
	"conv_expand": 2,
	"dropout": 0.0,
	"dtype": "bfloat16",
	"eos_token_id": 32009,
	"fusion": "gated",
	"gated_delta_conv_kernel_size": 4,
	"gated_delta_every": 0,
	"gated_delta_gate_init": 0.0,
	"gated_delta_key_dim": 64,
	"gated_delta_num_heads": 4,
	"gated_delta_start_layer": 0,
	"gated_delta_value_dim": 64,
	"grid_size": 256,
	"hidden_act": "silu",
	"hidden_size": 512,
	"initializer_range": 0.02,
	"intermediate_size": 2048,
	"linear_attention_every": 0,
	"linear_attention_gate_init": 0.0,
	"linear_attention_num_heads": 4,
	"linear_attention_start_layer": 0,
	"max_grid_tokens": 65536,
	"max_position_embeddings": 65536,
	"model_type": "convgpt_v2",
	"num_hidden_layers": 32,
	"pack_order": "sequence_to_curve",
	"packing": "hilbert",
	"pad_token_id": 32000,
	"position_embedding_type": "rope_nope",
	"retrieval_dropout": 0.0,
	"retrieval_every": 2,
	"retrieval_num_heads": 4,
	"retrieval_num_slots": 64,
	"retrieval_top_k": 4,
	"rms_norm_eps": 1e-06,
	"rope_theta": 10000.0,
	"router_rope_fraction": 1.0,
	"router_type": "chunk_token_memory",
	"tie_word_embeddings": true,
	"transformers_version": "5.9.0",
	"two_d_every": 1,
	"two_d_start_layer": 0,
	"use_1d_branch": true,
	"use_2d_branch": true,
	"use_2d_depthwise": true,
	"use_cache": false,
	"use_row_col_embeddings": false,
	"vocab_size": 32024
	}