speed1313/jax-llm

Issues

• 古文オンリーデータセット

  #24 opened 8 months ago
  1

• os path join使う

  #23 opened 8 months ago
  0

• config file や file_pathの整理

  #22 opened 8 months ago
  0

• add decoding algo

  #21 opened 8 months ago
  0

• efficient generate

  #20 opened 8 months ago
  0

• 損失のところソフトマックスかけてるか見る

  #19 opened 8 months ago
  0

• dataloaderをJAX likeにする

  #18 opened 8 months ago
  0

• dropout 0にしてみる

  #17 opened 8 months ago
  0

• Shakespeare で損失の下がり具合見る

  #16 opened 8 months ago
  0

• mhaを消してみる

  #15 opened 8 months ago
  1

• wikitext-ja

  #14 opened 8 months ago
  1

• Publish models on HuggingFace

  #13 opened 8 months ago
  0

• Training on 青空文庫

  #12 opened 8 months ago
  1

• training on my blog posts

  #11 opened 8 months ago
  0

• get batches with dynamic slice

  #10 opened 8 months ago
  1

• Dataclass GPTConfig

  #9 opened 8 months ago
  0

• Check if Token embedding indicates the nature of word embedding

  #8 opened 8 months ago
  1

• Code reading about Tiktoken

  #7 opened 8 months ago
  0

• RLHF

  #6 opened 8 months ago
  0

• fine tuning

  #5 opened 8 months ago
  0

• vocab size from 50257 to 50304

  #4 opened 8 months ago
  0

• distributed training

  #3 opened 8 months ago
  2

• evaluate my llm

  #2 opened 8 months ago
  0

• explore why my gpt-2 model's total params is about 80M

  #1 opened 8 months ago
  1