Datasets:

CMSManhattan
/

JiRack-Pretrain-Dataset

Tasks:

Text Generation

Modalities:

Formats:

Languages:

Size:

Tags:

Libraries:

License:

Dataset card Data Studio Files Files and versions

kgrabko commited on Jul 3

Commit

67dab4c

·

verified ·

1 Parent(s): f7ded37

Update README.md

Files changed (1) hide show

README.md +6 -1

README.md CHANGED Viewed

@@ -26,7 +26,12 @@ So JiRack Tokenizers stand out because they’ve been trained on premium **JiRac
 1. JiRack-Pro-Tokenizer
 https://huggingface.co/CMSManhattan/JiRack-Pro-Tokenizer-128K
-2. JiRack-Router-Tokenizer
 https://huggingface.co/CMSManhattan/JiRack-Router-Tokenizer-65K

 1. JiRack-Pro-Tokenizer
 https://huggingface.co/CMSManhattan/JiRack-Pro-Tokenizer-128K
+```bash
+python train_jirack_accelerate.py
+Processing jirack_pretrain_chunk_0.pt:  27%|█████████████████████▋                                                  | 268/1000 [22:02:45<60:12:33, 296.11s/it, loss=6.3145, avg_loss=7.0132, ppl=1111.16]
+Processing jirack_pretrain_chunk_0.pt:  87%|███████████████████████████████████████████████████████████████████████      | 866/1000 [75:57:47<13:13:29, 355.29s/it, loss=2.8616, avg_loss=5.9877, ppl=398.52]
+3. JiRack-Router-Tokenizer
 https://huggingface.co/CMSManhattan/JiRack-Router-Tokenizer-65K