IlyaGusev
/

llama_13b_ru_turbo_alpaca_lora

Text2Text Generation

Model card Files Files and versions Community

IlyaGusev commited on Apr 8, 2023

Commit

dfc2610

•

1 Parent(s): 0123035

Update README.md

Files changed (1) hide show

README.md +4 -2

README.md CHANGED Viewed

@@ -20,12 +20,14 @@ Colab: [link](https://colab.research.google.com/drive/18GG-zwxyGAHoVXghlOQONkxb_
 Training code: [link](https://github.com/IlyaGusev/rulm/tree/master/self_instruct)
 Versions:
-- v2, April 8, 2023: A model trained with outputs capped at 800 tokens instead of 400. `revision="3dff86a20f80ba203fec6e986a3422b83822ca70"`
   - Final loss: 0.995
   - Time: 38 hours, 35 minutes
   - Hardware: 1x RTX 3090Ti
   - Wandb link: [link](https://wandb.ai/ilyagusev/rulm_self_instruct/runs/wy66bwim/overview?workspace=user-ilyagusev)
-- v1, April 2, 2023: Original model trained with 400 max tokens in outputs. Both Russian and English are supported, trained on the ru_turbo_alpaca and alpaca-cleaned. `revision="97209c36c9261eb9527371a341c70f90c39ee158"`
   - Final loss: 1.074
   - Time: 53 hours, 53 minutes
   - Hardware: 1x RTX 3090

 Training code: [link](https://github.com/IlyaGusev/rulm/tree/master/self_instruct)
 Versions:
+- v2, April 8, 2023: A model trained with outputs capped at 800 tokens instead of 400.
+  - `revision="3dff86a20f80ba203fec6e986a3422b83822ca70"`
   - Final loss: 0.995
   - Time: 38 hours, 35 minutes
   - Hardware: 1x RTX 3090Ti
   - Wandb link: [link](https://wandb.ai/ilyagusev/rulm_self_instruct/runs/wy66bwim/overview?workspace=user-ilyagusev)
+- v1, April 2, 2023: Original model trained with 400 max tokens in outputs. Both Russian and English are supported, trained on the ru_turbo_alpaca and alpaca-cleaned.
+  - `revision="97209c36c9261eb9527371a341c70f90c39ee158"`
   - Final loss: 1.074
   - Time: 53 hours, 53 minutes
   - Hardware: 1x RTX 3090