Gavi Narra

One observation with custom data, make sure the text is less than 2048 tokens length in train, test and valid jsonl files. I wrote a quick program to check and...

Before fix on M3 Max (40 core GPU) After fix

I am not sure what to trust though. If you look at the plist file that is generated by `powermetrics` , the p-cores were showing usage but like you said...

@phgrosjean I think your change is much better. I am updating my PR

I updated the PR

wanted to chime in - same issue M3 Max 40 CPU

Seems like line 88 in `parsers.py` the idle ratio for cluster is showing as 0. This causes this line ```python cpu_metric_dict[name+"_active"] = int((1 - cluster["idle_ratio"])*100) ``` to show 100 for...

Just created #68

In my case the download_complete was getting the model_name and model_base with html tags, I cleaned up with the following ```python print(f"model_name: {model_name}") print(f"model_base: {model_base}") model_name = model_name.replace("","").replace("","").strip() model_base =...

any updates? I still seem to get this error llama.cpp/ggml.o: No such file or directory