chess-littletestmodel

chess-littletestmodel

Chess model for LLM Course Challenge.

• By: MDaytek
• Params: 790,560
• Architecture: GPT-2 with custom tokenizer

Usage

The model uses a custom tokenizer. Load it with:

from transformers import GPT2LMHeadModel, AutoConfig
import json

config = AutoConfig.from_pretrained("LLM-course/chess-littletestmodel")
model = GPT2LMHeadModel.from_pretrained("LLM-course/chess-littletestmodel", config=config)

# Load vocab
with open("vocab.json") as f:
    vocab = json.load(f)