AnirudhRajagopalan1201/tinystories-custom-20M

---

Model trained on the TinyStories Dataset, replicating https://arxiv.org/abs/2305.07759, based on GPT-Neo architecture.

---

Hyperparams used to train this model: "batch_size": 32, "block_size": 256, "lr": 5e-4, "n_layer": 6, "n_head": 6, "n_embd": 288, "dropout": 0.1, "weight_decay": 0.01, "epochs": 2, "eval_interval": 200, "eval_steps": 50, "vocab_size": 50257, "warmup_tokens": 10000, "gradient_accumulation_steps": 16

EXAMPLE USAGE

  !pip install --quiet transformers 
  from transformers import AutoModelForCausalLM, AutoTokenizer
  model20 = AutoModelForCausalLM.from_pretrained('AnirudhRajagopalan1201/tinystories-custom-20M')
  
  tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neo-125M")
  prompt = "Lily likes cats and dogs. She asked her mom for a dog and her mom said no, so instead she asked"
  input_ids = tokenizer.encode(prompt, return_tensors="pt") 
  output = model20.generate(input_ids, temperature=0.2, max_length = 100, do_sample=True)
  output_text = tokenizer.decode(output[0], skip_special_tokens=True)
  print(output_text)