Instructions to use VoCuc/anti-distill with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use VoCuc/anti-distill with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-R1-Distill-Qwen-7B") model = PeftModel.from_pretrained(base_model, "VoCuc/anti-distill") - Transformers
How to use VoCuc/anti-distill with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="VoCuc/anti-distill") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("VoCuc/anti-distill", dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use VoCuc/anti-distill with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "VoCuc/anti-distill" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "VoCuc/anti-distill", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/VoCuc/anti-distill
- SGLang
How to use VoCuc/anti-distill 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 "VoCuc/anti-distill" \ --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": "VoCuc/anti-distill", "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 "VoCuc/anti-distill" \ --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": "VoCuc/anti-distill", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use VoCuc/anti-distill with Docker Model Runner:
docker model run hf.co/VoCuc/anti-distill
| {"model_path": "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "ckpt_name": "DeepSeek", "model_type": "qwen", "student_model_type": null, "n_gpu": 1, "n_nodes": 1, "student_model_path": "Qwen/Qwen2.5-1.5B", "student_ckpt_name": "Qwen", "student_model_fp16": true, "model_parallel": false, "model_parallel_size": null, "no_value": false, "dropout_path_rate": null, "fp32": false, "bf16": false, "type": "fkl", "do_train": true, "do_valid": true, "do_eval": false, "base_path": ".", "load": null, "save": "./results/DeepSeek-R1-Distill-Qwen-7B", "log_interval": 20, "mid_log_num": -1, "save_interval": 600, "eval_interval": -1, "local_rank": 0, "save_additional_suffix": "", "save_rollout": false, "eb_sample_times": 3, "data_dir": "./processed_data/MetaMathQA-50k/qwen/", "processed_data_dir": null, "force_process": false, "force_process_demo": false, "data_process_workers": -1, "train_num": -1, "train_ratio": 1, "dev_num": 1000, "dev_ratio": 1, "gen_num": -1, "data_names": null, "prompt_type": null, "num_workers": 1, "max_prompt_length": 128, "min_prompt_length": 128, "json_data": false, "bin_data": false, "txt_data": false, "prompt_data_dir": null, "lm_data_dir": null, "eval_ppl": false, "eval_rw": false, "eval_gen": true, "only_prompt": false, "batch_size": 6, "eval_batch_size": 64, "clip_grad": 1.0, "total_iters": null, "train_iters_per_epoch": -1, "max_length": 600, "seed": 42, "seed_order": 42, "seed_data": 42, "seed_ppo": 42, "seed_lm": 7, "epochs": 5, "training_epochs": 10000, "gradient_accumulation_steps": 4, "gradient_checkpointing": false, "attn_dtype": null, "lr": 0.0003, "lr_min": 1e-07, "weight_decay": 0.0, "loss_scale": 65536, "kd_ratio": 1.0, "warmup_iters": 0, "warmup_ratio": 0.05, "lr_decay_iters": null, "lr_decay_style": "wrmup_cosine", "scheduler_name": "constant_trm", "w_span_loss": 1.0, "reward_scaling": null, "cliprange_reward": 1, "ppo_epochs": null, "num_rollouts": 256, "num_rollouts_per_device": null, "cliprange": 0.2, "chunk_size": null, "gamma": 0.95, "skew_alpha": 0.1, "student_gen": false, "gen_top_p": 1.0, "gen_num_beams": 2, "mixed_alpha": 0.5, "loss_eps": 0.1, "init_threshold": 0.0, "capacity": 1000, "replay_ratio": "decreasing", "student_lr": 0.01, "delta_lamda": 0.1, "top_k": 0, "top_p": 1.0, "do_sample": true, "no_repeat_ngram_size": 6, "repetition_penalty": null, "num_beams": 1, "temperature": 0.7, "peft": null, "peft_lora_r": 16, "peft_lora_alpha": 32, "peft_lora_dropout": 0.05, "peft_name": null, "peft_path": null, "student_peft_name": null, "student_peft_path": null, "deepspeed": true, "deepspeed_config": "./configs/deepspeed/ds_config_bf16.json", "deepscale": false, "deepscale_config": null, "rank": 0, "world_size": 1} |