Yewen Zhou
Yewen Zhou
I see. By managing multiple workers, does it have to work given that the model itself has the capacity to support multiple workers?
Thanks! Just to confirm again, by multiple workers, do you mean multiple GPUs?
I am still confused - maybe because I am not very versed in Java. So if I understand you correctly, `Cumulative inference duration` should be the largest, right? Because it...
Thank you so much for the comment! I tried re-writing my code using `asyncio` and `aiohttp` and the python file looks like below ``` import aiohttp import asyncio import time...
+1, what is the license for the pretrained models? On Hugging Face they are license under Apache-2.0. However, if they are fine-tuned with Alpaca, technically it shouldn't be licensed under...
> [PP-OCRv4 arabic model](https://github.com/PaddlePaddle/PaddleOCR/blob/dygraph/paddleocr.py#L135C27-L135C27) This is an inference model, what about pre-trained model for Arabic?
Is the dictionary same?
thanks @TheSeriousProgrammer and @thinh-huynh-re for your replies. It turned out that the real issue is #192. After I downgraded both timm and transformers versions, the results made much more sense....
If I remember I was able to solve this issue by downgrading the packages following #192
Could be related to this https://github.com/clovaai/donut/issues/205? Already submitted a PR https://github.com/clovaai/synthtiger/pull/58