Mofi Rahman

I will get to it once I am back from vacation next week. Thanks for the heads up. On Thu, Mar 28, 2024 at 3:36 PM Aldo Culquicondor ***@***.***> wrote:...

@alculquicondor

/assign

``` trainer.save_model(new_model_id) # Reload model in FP16 and merge it with LoRA weights base_model = AutoModelForCausalLM.from_pretrained( model_id, low_cpu_mem_usage=True, return_dict=True, torch_dtype=torch.bfloat16, ) newmodel = PeftModel.from_pretrained(base_model, new_model_id) newmodel = newmodel.merge_and_unload() print(check_model_weights_equality(model, newmodel))...

I can take a look.

Followed the instruction from @tjungblu and managed to reproduce the error. `client/pkg/testutil/recorder.go` L131 has a `case

@ahrtr I would like to work on it more. But I am not sure what would be the path to solving this. What should I look into more?

I would like to take a stab at it.

/assign

@SlashmanX Your answer finally solved few hours of work. I just have few more questions. trakt_client_id, trakt_client_secret and trakt_redirect_uri is something I would get from trakt api? Also TraktoidProvider.AUTHORITY does...