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

Libraries

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

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

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

# pip install -U transformers accelerate
# Load model directly
from transformers import LongcatCausalLM
model = LongcatCausalLM.from_pretrained("tiny-random/longcat-flash-lite-sparse", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

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

Use Docker

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

SGLang

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

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

yujiepan commited on 1 day ago

Commit

462e5ce

verified ·

1 Parent(s): 42072a7

Upload folder using huggingface_hub

Browse files

Files changed (2) hide show

.meta/model.log +1 -0
README.md +7 -3

.meta/model.log CHANGED Viewed

	@@ -0,0 +1 @@

+ LongcatCausalLM (direct SGLang-schema tiny checkpoint): 4 dual-sublayer LongCat Flash decoder layers with shortcut MoE (32 routed + 128 identity zero experts, top-k 12), compressed MLA (q_lora_rank=32, kv_lora_rank=32, 32 heads, nope 32 / rope 64 / value 32), LSA indexers (16 heads, head dim 128, CLI factor 2), two over-embedding windows, one MTP/NextN draft layer, 131,072-row untied token tables.

README.md CHANGED Viewed

@@ -25,11 +25,15 @@ This is a randomly initialized tiny model based on [meituan-longcat/LongCat-Flas
 | Routers / correction bias | FP32 | Same |
 | Stored precision | BF16 | Same |
-The source repo ships no model code and no Transformers implementation (`config.json` has no `model_type` or `auto_map`); SGLang recognizes its `LongcatCausalLM` architecture through its `longcat_flash` config parser and DSA indexer. The tiny tensors were generated directly against that loader's state-dict schema and key remapping, including the nested `model.ngram_embeddings.{embedders,post_projs}.*` over-embedding names (newer SGLang also accepts the source checkpoint's flattened `oe_embed_tokens*.weight` aliases) and the `model.mtp.*` NextN draft. The n-gram hash modulus is internal embedding machinery rather than an externally visible task space, so it follows the smaller [tiny-random/longcat-flash-lite](https://huggingface.co/tiny-random/longcat-flash-lite) modulus; the 131,072 task vocabulary rows and token ids are unchanged. No local Transformers execution path exists; serve with SGLang (0.5.23.dev384 or newer, which includes LongCat LSA) using the command below.
 ### Example usage
-#### SGLang serving (✅ verified)
 LSA is implemented in SGLang (main branch; the LongCat Sparse adaptation landed via PR #32918). The checkpoint keeps the source's LongcatCausalLM architecture and serves text only, like the upstream deployment guide.
@@ -59,6 +63,6 @@ curl http://127.0.0.1:30000/v1/chat/completions \
 ### Change log
-- 2026-11-03: Initial version. Direct SGLang-schema checkpoint: Transformers has no implementation for the LongCat Sparse Attention (LSA) repo (no model_type/auto_map), so tensors follow the pinned sgl-project/sglang main loader (`longcat_flash.py` plus the DSA `Indexer`). Four dual-sublayer logical layers (eight attention sublayers) keep the shortcut-MoE topology with 32 routed and 128 weightless identity zero experts at top-k 12, compressed MLA ranks (q 32 / kv 256, the Hopper FA3 compressed-latent q_v kernel floor; the source uses kv 512) with 8 Q heads, 64/64 nope/rope and the source 128 value head dim (qk 128 + v 128 matches the Hopper FA3 MHA/MLA head-dim windows), and LSA indexers with the source 16 heads / head dim 128 on every even physical attention layer (CLI factor 2). One MTP/NextN draft layer with its own indexer is serialized under `model.mtp.*`; over-embeddings use the loader-native nested ngram names and the sibling tiny modulus (ratio 4, 2 x 2 windows) instead of the source 10.2-million-row hash tables.
 *This repo is generated by GPT-6.1-Sol.*

 | Routers / correction bias | FP32 | Same |
 | Stored precision | BF16 | Same |
+SGLang-only: the source ships no Transformers code and the tensors follow SGLang's `longcat_flash` schema. The n-gram hash modulus is internal embedding machinery, so it follows [tiny-random/longcat-flash-lite](https://huggingface.co/tiny-random/longcat-flash-lite).
+| File path | Size |
+|------|------|
+| model.safetensors | 23.5MB |
 ### Example usage
+#### SGLang serving (unverified)
 LSA is implemented in SGLang (main branch; the LongCat Sparse adaptation landed via PR #32918). The checkpoint keeps the source's LongcatCausalLM architecture and serves text only, like the upstream deployment guide.
 ### Change log
+- 2026-11-03: Initial version: direct SGLang-schema LSA checkpoint (shortcut-MoE, CLI indexers, ngram over-embeddings and one NextN MTP draft).
 *This repo is generated by GPT-6.1-Sol.*