Text Generation
Transformers
Safetensors
English
fwkv
feature-extraction
rnn
conversational
custom_code
Instructions to use FWKV/Fables with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use FWKV/Fables with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="FWKV/Fables", trust_remote_code=True) messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("FWKV/Fables", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use FWKV/Fables with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "FWKV/Fables" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "FWKV/Fables", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/FWKV/Fables
- SGLang
How to use FWKV/Fables 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 "FWKV/Fables" \ --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": "FWKV/Fables", "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 "FWKV/Fables" \ --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": "FWKV/Fables", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use FWKV/Fables with Docker Model Runner:
docker model run hf.co/FWKV/Fables
Download config.json from FWKV/Fables: direct link, hf CLI and curl.
- Browser
- Download file 1.24 kB
-
https://huggingface.co/FWKV/Fables/resolve/main/config.json
- Command line
-
hf download hf://FWKV/Fables/config.json
-
curl -L -o config.json https://huggingface.co/FWKV/Fables/resolve/main/config.json
1.24 kB
| { | |
| "accum_steps": 8, | |
| "arch_version": 2, | |
| "auto_map": { | |
| "AutoConfig": "configuration_fwkv.FWKVConfig", | |
| "AutoModel": "modeling_fwkv.FWKVLanguageModel", | |
| "AutoModelForCausalLM": "modeling_fwkv.FWKVLanguageModel" | |
| }, | |
| "architectures": [ | |
| "FWKVLanguageModel" | |
| ], | |
| "batch_size": 4, | |
| "ce_chunk": 512, | |
| "chinchilla_scale": true, | |
| "d_emb": 384, | |
| "d_model": 768, | |
| "dtype": "float32", | |
| "early_stop_patience": 2, | |
| "epochs": 1, | |
| "eval_every_opt_steps": 100, | |
| "ffn_mult": 4, | |
| "grad_clip": 1.0, | |
| "head_size": 64, | |
| "keep_last_ckpts": 2, | |
| "log_every": 50, | |
| "lr": 0.0004, | |
| "max_val_examples": 128, | |
| "min_lr_ratio": 0.1, | |
| "model_type": "fwkv", | |
| "n_layers": 13, | |
| "prefetch_chunks": 512, | |
| "save_every_steps": 500, | |
| "seq_len": 1024, | |
| "sft_epochs": 3, | |
| "sft_lr": 5e-05, | |
| "sft_max_train_examples": 40000, | |
| "sft_max_val_examples": 1000, | |
| "sft_val_skip_rows": 2000, | |
| "shuffle_buffer_size": 10000, | |
| "tau_max": 1000.0, | |
| "tau_min": 2.0, | |
| "tie_word_embeddings": true, | |
| "train_tokens": 20000000000, | |
| "transformers_version": "5.16.1", | |
| "use_amp": true, | |
| "vocab_size": 50280, | |
| "warmup_steps": 1000, | |
| "weight_decay": 0.1, | |
| "wikitext_max_chunks": 128, | |
| "wikitext_val": true, | |
| "wkv_ceil": 0.999, | |
| "wkv_floor": 0.1 | |
| } | |