Instructions to use tiny-random/longcat-flash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiny-random/longcat-flash with Transformers:

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

pipe = pipeline("text-generation", model="tiny-random/longcat-flash", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tiny-random/longcat-flash", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("tiny-random/longcat-flash", trust_remote_code=True)
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

vLLM

How to use tiny-random/longcat-flash with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiny-random/longcat-flash"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/longcat-flash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tiny-random/longcat-flash

SGLang

How to use tiny-random/longcat-flash 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 "tiny-random/longcat-flash" \
    --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": "tiny-random/longcat-flash",
		"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 "tiny-random/longcat-flash" \
        --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": "tiny-random/longcat-flash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tiny-random/longcat-flash with Docker Model Runner:
```
docker model run hf.co/tiny-random/longcat-flash
```

longcat-flash / config.json

yujiepan

Upload folder using huggingface_hub

35f4a5f verified 9 months ago

raw

history blame contribute delete

1.29 kB

	{
	"architectures": [
	"LongcatFlashForCausalLM"
	],
	"attention_bias": false,
	"attention_dropout": 0.0,
	"attention_method": "MLA",
	"auto_map": {
	"AutoConfig": "configuration_longcat_flash.LongcatFlashConfig",
	"AutoModel": "modeling_longcat_flash.LongcatFlashModel",
	"AutoModelForCausalLM": "modeling_longcat_flash.LongcatFlashForCausalLM"
	},
	"bos_token_id": 1,
	"eos_token_id": 2,
	"expert_ffn_hidden_size": 64,
	"ffn_hidden_size": 64,
	"head_dim": 192,
	"hidden_act": "silu",
	"hidden_size": 8,
	"initializer_range": 0.006,
	"kv_lora_rank": 384,
	"max_position_embeddings": 131072,
	"mla_scale_kv_lora": true,
	"mla_scale_q_lora": true,
	"model_type": "longcat_flash",
	"moe_topk": 12,
	"n_routed_experts": 32,
	"norm_topk_prob": false,
	"num_attention_heads": 4,
	"num_key_value_heads": 4,
	"num_layers": 2,
	"q_lora_rank": 32,
	"qk_head_dim": 256,
	"qk_nope_head_dim": 64,
	"qk_rope_head_dim": 192,
	"rms_norm_eps": 1e-05,
	"rope_theta": 10000000.0,
	"routed_scaling_factor": 6.0,
	"router_bias": false,
	"tie_word_embeddings": false,
	"torch_dtype": "bfloat16",
	"transformers_version": "4.56.0.dev0",
	"use_cache": true,
	"v_head_dim": 64,
	"vocab_size": 131072,
	"zero_expert_num": 16,
	"zero_expert_type": "identity"
	}