Neko-Institute-of-Science
commited on
Commit
•
01e308f
1
Parent(s):
1d30ac2
Training Done.
Browse files
README.md
CHANGED
@@ -17,12 +17,15 @@ Also I will be uploading checkpoints almost everyday. I could train another epoc
|
|
17 |
|
18 |
Update: Since I will not be training over 1 epoch @Aeala is training for the full 3 https://huggingface.co/Aeala/VicUnlocked-alpaca-half-30b-LoRA but it's half ctx if you care about that. Also @Aeala's just about done.
|
19 |
|
|
|
|
|
20 |
# How to test?
|
21 |
1. Download LLaMA-30B-HF if you have not: https://huggingface.co/Neko-Institute-of-Science/LLaMA-30B-HF
|
22 |
-
2.
|
23 |
-
3.
|
24 |
-
4.
|
|
|
25 |
|
26 |
|
27 |
-
#
|
28 |
https://wandb.ai/neko-science/VicUnLocked/runs/vx8yzwi7
|
|
|
17 |
|
18 |
Update: Since I will not be training over 1 epoch @Aeala is training for the full 3 https://huggingface.co/Aeala/VicUnlocked-alpaca-half-30b-LoRA but it's half ctx if you care about that. Also @Aeala's just about done.
|
19 |
|
20 |
+
Update: Training Finished at Epoch 1, These 8 days sure felt long. I only have one A6000 lads there's only so much I can do. Also RIP gozfarb IDK what happened to him.
|
21 |
+
|
22 |
# How to test?
|
23 |
1. Download LLaMA-30B-HF if you have not: https://huggingface.co/Neko-Institute-of-Science/LLaMA-30B-HF
|
24 |
+
2. Make a folder called VicUnLocked-30b-LoRA in the loras folder.
|
25 |
+
3. Download adapter_config.json and adapter_model.bin into VicUnLocked-30b-LoRA.
|
26 |
+
4. Load ooba: ```python server.py --listen --model LLaMA-30B-HF --load-in-8bit --chat --lora VicUnLocked-30b-LoRA```
|
27 |
+
5. Select instruct and chose Vicuna-v1.1 template.
|
28 |
|
29 |
|
30 |
+
# Training Log
|
31 |
https://wandb.ai/neko-science/VicUnLocked/runs/vx8yzwi7
|