Image-Text-to-Text
Transformers
Safetensors
qwen3_5
multimodal
visual-token-compression
on-policy-distillation
conversational
Instructions to use yyy051007/LT-OPD with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use yyy051007/LT-OPD with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="yyy051007/LT-OPD") 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("yyy051007/LT-OPD") model = AutoModelForMultimodalLM.from_pretrained("yyy051007/LT-OPD", 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 yyy051007/LT-OPD with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "yyy051007/LT-OPD" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "yyy051007/LT-OPD", "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/yyy051007/LT-OPD
- SGLang
How to use yyy051007/LT-OPD 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 "yyy051007/LT-OPD" \ --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": "yyy051007/LT-OPD", "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 "yyy051007/LT-OPD" \ --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": "yyy051007/LT-OPD", "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 yyy051007/LT-OPD with Docker Model Runner:
docker model run hf.co/yyy051007/LT-OPD
Download visual_compression_config.json from yyy051007/LT-OPD: direct link, hf CLI and curl.
- Browser
- Download file 3.02 kB
-
https://huggingface.co/yyy051007/LT-OPD/resolve/main/visual_compression_config.json
- Command line
-
hf download hf://yyy051007/LT-OPD/visual_compression_config.json
-
curl -L -o visual_compression_config.json https://huggingface.co/yyy051007/LT-OPD/resolve/main/visual_compression_config.json
3.02 kB
| { | |
| "attn_implementation": "flash_attention_2", | |
| "vision_token_compressor": { | |
| "enabled": true, | |
| "algorithm": "qwen35_cdpruner_v1", | |
| "method": "cdpruner", | |
| "paper_url": "https://arxiv.org/abs/2506.10967", | |
| "official_repo": "Theia-4869/CDPruner", | |
| "official_repo_revision": "9541616c40fcd5625de1cdb8ea6c33c129eb7864", | |
| "adaptation_scope": "qwen35_single_image_post_native_merger_cdpruner_learnable_nearest_retained_merge_pre_llm", | |
| "original_method_reproduction_claim_allowed": false, | |
| "route_schema_version": "vision_opd_cdpruner_learnable_merge_zero_residual_curriculum_route_v1", | |
| "transport_key": "dart_merge_routes", | |
| "placement": "post_native_merger_pre_llm", | |
| "budget_policy": "per_image_min_N_max_32_ceil_active_optimizer_step_retention_bps_N_v1", | |
| "retention_ratio": 0.05, | |
| "minimum_tokens_per_image": 32, | |
| "distance": null, | |
| "distance_chunk_tokens": null, | |
| "routing_autograd": "detached_no_grad_fp32", | |
| "knn_k": null, | |
| "center_score": null, | |
| "assignment": "greedy_conditional_dpp_stable_original_index_tie", | |
| "reduction": "differentiable_fp32_zero_init_signed_residual", | |
| "position_policy": "selected_original_mrope", | |
| "query_conditioned": true, | |
| "attention_conditioned": false, | |
| "bbox_conditioned": false, | |
| "persistent_n_by_n_matrix_forbidden": true, | |
| "kernel_jitter": 1e-06, | |
| "relevance_epsilon": 1e-06, | |
| "residual_epsilon": 1e-12, | |
| "selector_dtype": "float32", | |
| "merge": "learnable_nearest_retained_zero_init_residual", | |
| "query_policy": "user_question_options_only_excluding_special_and_image_tokens", | |
| "relevance_fraction": null, | |
| "candidate_multiplier": null, | |
| "query_weight": null, | |
| "attention_weight": null, | |
| "bbox_weight": null, | |
| "deterministic": true, | |
| "data_manifest_path": null, | |
| "data_manifest_schema": null, | |
| "conditional_objective": "instruction_relevance_weighted_visual_diversity_greedy_dpp_matrix_free_fp32", | |
| "assignment_chunk_tokens": 2048, | |
| "query_route_schema_version": "vision_opd_route_query_v2", | |
| "summary": { | |
| "enabled": false | |
| }, | |
| "curriculum": { | |
| "enabled": true, | |
| "schema_version": "vision_opd_v8_visual_token_curriculum_v2", | |
| "driver": "completed_optimizer_steps_before_current_update", | |
| "schedule_type": "warmup_hold_raised_cosine_frozen_bps", | |
| "total_optimizer_steps": 188, | |
| "start_retention_bps": 2500, | |
| "final_retention_bps": 500, | |
| "warmup_hold_optimizer_steps": 14, | |
| "final_plateau_start_completed_steps": 100, | |
| "rounding_mode": "round_half_up", | |
| "minimum_tokens_per_image": 32 | |
| }, | |
| "learnable_merge": { | |
| "enabled": true, | |
| "aggregation": "zero_init_signed_residual_v1", | |
| "initialization": "pretrained_mlp_gate_rows_zero_output_v1", | |
| "input_hidden_size": 2560, | |
| "hidden_size": 64, | |
| "temperature": 1.0, | |
| "assignment": "cosine_nearest_retained_original_index_tie", | |
| "activation_checkpointing": true, | |
| "chunk_tokens": 1024 | |
| } | |
| } | |
| } | |