-
Currently, the community has started experimenting with building more models using a mix of different local experts. In the current implementation of mlx-lm, we have hardcoded the linear_class_predica…
mzbac updated
4 months ago
-
src/loragw_hal.c:260:14: warning: taking the absolute value of unsigned type ‘uint32_t’ {aka ‘unsigned int’} has no effect [-Wabsolute-value]
https://github.com/Lora-net/sx1302_hal/blob/4b42025d175…
-
- sd1.5
- It works when resume is removed.
- Using network_weight instead works.
- If I comment out network_args and create resume and load it, it works fine (algo=full specific problem?).
- I do …
ghost updated
6 months ago
-
Hi there,
Could it be that by disabling the lora-network-server we also did something with the maxTxPower? Since running the installer script on a spare Conduit I lost a lot of coverage with TTN ma…
-
Hello, thank you for sharing the source code. While trying to reproduce **SST2 task result with RoBERTa-base model**, I've encountered some questions regarding the hyper-parameters, lora_alpha, and a …
-
The training benchmark link no longer works : https://huggingface.co/blog/huggingface-and-optimum-amd
How can one test training throughput on AMD these days?
Also, can you provide details about th…
-
hello @rpp0 @riatankarsahu
I have seen your problem ModuleNotFoundError: No module named 'lora' ,I am experiencing the same problem after uploading gr-lora in to GNURadio ,
1)I did the librarypath …
-
### Description
When running the mlx-usft.ipynb notebook on M1 Mac with the `--adapter-file` argument, it results in an "unrecognized arguments" error. It seems like the argument is either not imple…
-
really cool project! im wondering how its different from s-Lora? https://github.com/S-LoRA/S-LoRA
-
Based on the training speed, sample generation speed, and final file size compared to Diag-OFT, it appears that BOFT-training has switched to float32. Is there a possibility to implement mixed precisi…
iqddd updated
2 months ago