Mirrors/text-generation-webui
1
0
Fork 0
mirror of https://github.com/oobabooga/text-generation-webui.git synced 2024-11-22 08:07:56 +01:00
Code Issues Packages Projects Releases Wiki Activity

Update Using-LoRAs.md

Browse Source
This commit is contained in:
oobabooga 2023-04-22 02:53:01 -03:00 committed by GitHub
parent 6d4f131d0a
commit 9508f207ba
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
docs/Using-LoRAs.md
View File

@ -16,7 +16,7 @@ python server.py --model llama-7b-hf --lora alpaca-lora-7b --load-in-8bit
python server.py --model llama-7b-hf --lora alpaca-lora-7b --cpu python server.py --model llama-7b-hf --lora alpaca-lora-7b --cpu
``` ```
* For using LoRAs in 4-bit mode, follow these special instructions: https://github.com/oobabooga/text-generation-webui/wiki/GPTQ-models-(4-bit-mode)#using-loras-in-4-bit-mode * For using LoRAs in 4-bit mode, follow [these special instructions](GPTQ-models-(4-bit-mode).md#using-loras-in-4-bit-mode).
* Instead of using the `--lora` command-line flag, you can also select the LoRA in the "Parameters" tab of the interface. * Instead of using the `--lora` command-line flag, you can also select the LoRA in the "Parameters" tab of the interface.
@ -85,4 +85,4 @@ Run the script with:
python finetune.py python finetune.py
``` ```
It just works. It runs at 22.32s/it, with 1170 iterations in total, so about 7 hours and a half for training a LoRA. RTX 3090, 18153MiB VRAM used, drawing maximum power (350W, room heater mode). It just works. It runs at 22.32s/it, with 1170 iterations in total, so about 7 hours and a half for training a LoRA. RTX 3090, 18153MiB VRAM used, drawing maximum power (350W, room heater mode).