MagicSource

By using xformers to train llama2, the loss are explosion, do u know why? On V100 only

About the print message

2

Hi, the printed message I can not understand. I set ``` def __init__( self, dim, ratio, max_position_embeddings=2048, base=10000, device=None ): ``` ratio=2, max_position_embeddings=1024 Since my GPU can not fit minimal...

About the learning rate

1

from the script provided, I think longchat is full sft rather than lora, but the equal batch size total is just 1 (batch_size * gradient_accum * num_gpus) But vicuna original...

Hoping for support training on some models that customized not inside transformers like baicuhan

Fixed a dummy url in source code?

Hi how to specific url path for custom openai like server?

Hi, this lib is useful. I want customized the link pattern, is there a way to do it? am using this logic to buld spans: ```dart class WidgetSpanTextEditingController extends TextEditingController...

this lib is not maintained, please switch to pytorch3d

Which data make the model learned table to markdown?

3

These opened dataset can not really find which dataset can hav img -> markdown text information. And where does the Chinese OCR ability comes from? The whole dataset has no...

windows build faild while llama-cpp-py worksa

1

windows build faild while llama-cpp-py worksa