Text Generation
Transformers
Safetensors
English
llama
abliterated
uncensored
refusal-removal
interpretability
senbonzakura
conversational
text-generation-inference
Instructions to use ElementMerc/tinyllama-1.1b-uncensored with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ElementMerc/tinyllama-1.1b-uncensored with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ElementMerc/tinyllama-1.1b-uncensored") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("ElementMerc/tinyllama-1.1b-uncensored") model = AutoModelForCausalLM.from_pretrained("ElementMerc/tinyllama-1.1b-uncensored", device_map="auto") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use ElementMerc/tinyllama-1.1b-uncensored with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ElementMerc/tinyllama-1.1b-uncensored" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ElementMerc/tinyllama-1.1b-uncensored", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/ElementMerc/tinyllama-1.1b-uncensored
- SGLang
How to use ElementMerc/tinyllama-1.1b-uncensored 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 "ElementMerc/tinyllama-1.1b-uncensored" \ --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": "ElementMerc/tinyllama-1.1b-uncensored", "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 "ElementMerc/tinyllama-1.1b-uncensored" \ --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": "ElementMerc/tinyllama-1.1b-uncensored", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use ElementMerc/tinyllama-1.1b-uncensored with Docker Model Runner:
docker model run hf.co/ElementMerc/tinyllama-1.1b-uncensored
Commit History
21 days ago
-
Rename the references to match the repository's new name
Verified signatureThe repositories were renamed from -abliterated to -uncensored. The Hub redirects the old names, so nothing was broken, but the ids inside the card are what people copy and run, and the GGUF cards' base_model field is what builds the quantized-from relation on the Hub. "Abliteration" is kept where it names the method, because that is what the method is called and the card explains it. Only the repository names changed.45e20a6
22 days ago
-
Point every reference at ElementMerc, the account that hosts these weights
Verified signatureThe account was renamed from ops-malware, and the Hub's redirects kept the old links working. A redirect is the platform's courtesy rather than a guarantee, and these references sit inside commands people copy and run, so they now name the account directly. The GGUF cards' base_model metadata is updated too, because that is the field the Hub reads to build the quantized-from relation, and it pointed at a name that no longer exists as an author. No weights, no figures and no correction notices are changed by this.94633b7
-
correction notice: every evaluation figure on this card is withdrawn
MercuryVerified signatureb4cd412
-
Add card: evaluation, limitations, licence
MercuryVerified signature375b180 -
Add card: evaluation, limitations, licence
MercuryVerified signature2021172 -
Upload folder using huggingface_hub
MercuryVerified signature98c8c04 -
initial commit
MercuryVerified signaturea8b9aab