Image-Text-to-Text
Transformers
Safetensors
English
Chinese
qwen3_5_moe
code
agent
agentic-coding
Mixture of Experts
vision-language
vllm
conversational
Instructions to use beyoru/KAT-Coder-V2.5-Dev-VL with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use beyoru/KAT-Coder-V2.5-Dev-VL with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="beyoru/KAT-Coder-V2.5-Dev-VL") 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)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("beyoru/KAT-Coder-V2.5-Dev-VL") model = AutoModelForMultimodalLM.from_pretrained("beyoru/KAT-Coder-V2.5-Dev-VL", device_map="auto") 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=256) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use beyoru/KAT-Coder-V2.5-Dev-VL with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "beyoru/KAT-Coder-V2.5-Dev-VL" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "beyoru/KAT-Coder-V2.5-Dev-VL", "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/beyoru/KAT-Coder-V2.5-Dev-VL
- SGLang
How to use beyoru/KAT-Coder-V2.5-Dev-VL 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 "beyoru/KAT-Coder-V2.5-Dev-VL" \ --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": "beyoru/KAT-Coder-V2.5-Dev-VL", "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 "beyoru/KAT-Coder-V2.5-Dev-VL" \ --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": "beyoru/KAT-Coder-V2.5-Dev-VL", "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 beyoru/KAT-Coder-V2.5-Dev-VL with Docker Model Runner:
docker model run hf.co/beyoru/KAT-Coder-V2.5-Dev-VL
Add docker-compose.vllm.yaml: sample serving config
Browse files- docker-compose.vllm.yaml +13 -0
docker-compose.vllm.yaml
CHANGED
|
@@ -28,6 +28,7 @@ services:
|
|
| 28 |
--model /models/KAT-Coder-V2.5-Dev-VL
|
| 29 |
--served-model-name kat-vl
|
| 30 |
--max-model-len 32768
|
|
|
|
| 31 |
--gpu-memory-utilization 0.90
|
| 32 |
--trust-remote-code
|
| 33 |
|
|
@@ -38,3 +39,15 @@ services:
|
|
| 38 |
# KV cache o context day rat lon. Muon vuot native thi ap YaRN y het huong dan cua
|
| 39 |
# Qwen/Qwen3.6-35B-A3B (tham so RoPE o day khong doi so voi upstream) — luu y moi
|
| 40 |
# framework deu cai YaRN TINH nen bat len se lam giam chat luong prompt ngan.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
--model /models/KAT-Coder-V2.5-Dev-VL
|
| 29 |
--served-model-name kat-vl
|
| 30 |
--max-model-len 32768
|
| 31 |
+
--max-num-batched-tokens 8192
|
| 32 |
--gpu-memory-utilization 0.90
|
| 33 |
--trust-remote-code
|
| 34 |
|
|
|
|
| 39 |
# KV cache o context day rat lon. Muon vuot native thi ap YaRN y het huong dan cua
|
| 40 |
# Qwen/Qwen3.6-35B-A3B (tham so RoPE o day khong doi so voi upstream) — luu y moi
|
| 41 |
# framework deu cai YaRN TINH nen bat len se lam giam chat luong prompt ngan.
|
| 42 |
+
|
| 43 |
+
# ⚠️ `--max-num-batched-tokens 8192` la BAT BUOC, dung bo di.
|
| 44 |
+
#
|
| 45 |
+
# Kien truc nay lai attention + Gated DeltaNet. Khi prefix caching bat (mac dinh o
|
| 46 |
+
# vLLM moi), vLLM ep `mamba_cache_mode='align'`, roi nang attention block size len
|
| 47 |
+
# 2096 de "attention page size >= mamba page size". Mac dinh max_num_batched_tokens
|
| 48 |
+
# chi 2048 => assert vo ngay luc khoi tao KV cache:
|
| 49 |
+
#
|
| 50 |
+
# AssertionError: In Mamba cache align mode,
|
| 51 |
+
# block_size (2096) must be <= max_num_batched_tokens (2048)
|
| 52 |
+
#
|
| 53 |
+
# Bat ky gia tri nao >= 2096 deu qua; 8192 con loi cho prefill.
|