You can deploy this model and start earning money today!

ljt019/
nanochat-d40

Name: nanochat-d40
Availability: InStock

Views

No views yet

nanochat d40 (SFT)

Custom nanochat checkpoint packaged for Hugging Face with trust_remote_code.

Usage

python

from transformers import AutoTokenizer, AutoModelForCausalLM

tok = AutoTokenizer.from_pretrained("ljt019/nanochat-d40", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("ljt019/nanochat-d40", trust_remote_code=True)

inputs = tok("Hello!", return_tensors="pt")
out = model.generate(**inputs, max_new_tokens=128, do_sample=True, temperature=0.6, top_k=50)
print(tok.decode(out[0].tolist()))

Notes

Weights are from chatsft_checkpoints/d40/model_000681.pt.
Tokenizer is loaded from tokenizer.pkl (tiktoken encoding).