GitBross Accueil Explorer Aide
S'inscrire Connexion
sol_87yf6cwp
/
LLMs-from-scratch
miroir de https://github.com/br8bit/LLMs-from-scratch.git
1
0
Fork 0
Fichiers Tickets 0 Wiki
Parcourir la source

update

rasbt il y a 1 an
Parent
6c0dc2362b
commit
4c2b503785
2 fichiers modifiรฉs avec 4 ajouts et 1 suppressions
Vue sรฉparรฉe Afficher les stats Diff
  1. 2 1
      ch07/01_main-chapter-code/README.md
  2. 2 0
      ch07/01_main-chapter-code/gpt_instruction_finetuning.py

+ 2 - 1
ch07/01_main-chapter-code/README.md
Voir le fichier 

@@ -30,6 +30,7 @@ Validation set length: 55
 
 Test set length: 110
 
 --------------------------------------------------
 
 Device: cpu
 
+--------------------------------------------------
 
 File already exists and is up-to-date: gpt2/355M/checkpoint
 
 File already exists and is up-to-date: gpt2/355M/encoder.json
 
 File already exists and is up-to-date: gpt2/355M/hparams.json
 
@@ -50,7 +51,7 @@ Training completed in 15.66 minutes.
 
 Plot saved as loss-plot-standalone.pdf
 
 --------------------------------------------------
 
 Generating responses
 
-100%|โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆ| 110/110 [06:57<00:00,  3.80s/it]
 
+100%|โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆ| 110/110 [06:57<00:00,  3.80s/it]
 
 Responses saved as instruction-data-with-response-standalone.json
 
 Model saved as gpt2-medium355M-sft-standalone.pth
 
 ```

+ 2 - 0
ch07/01_main-chapter-code/gpt_instruction_finetuning.py
Voir le fichier 

@@ -185,6 +185,8 @@ def main():
 
     tokenizer = tiktoken.get_encoding("gpt2")
 
     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
     print("Device:", device)
 
+    print(50*"-")
 
+
 
     customized_collate_fn = partial(custom_collate_fn, device=device, allowed_max_length=1024)
 
 
     num_workers = 0