Error merge_adapter on autotrain?

Hi everyone!

After few tests, i’m keeping an error during merging adapter to models.
I’m succefully train with a simple dataset (csv) based on Mistral models. After trying to merge it with the models, i got a failure with message: Connection errored out on the GUI.

Then i was tried to directly adjust the parameters and pass merge_adapter to “true”.

The training is ok, but i have an error about merging (seems to be).

Below, the parameters and the output log with error.

Did i miss something? Any ideas would be welcome :slight_smile:

Thx a lot

----Parameters----
{
“block_size”: 1024,
“model_max_length”: 2048,
“padding”: “right”,
“use_flash_attention_2”: false,
“disable_gradient_checkpointing”: false,
“logging_steps”: -1,
“evaluation_strategy”: “epoch”,
“save_total_limit”: 1,
“save_strategy”: “epoch”,
“auto_find_batch_size”: false,
“mixed_precision”: “fp16”,
“lr”: 0.00003,
“epochs”: 3,
“batch_size”: 2,
“warmup_ratio”: 0.1,
“gradient_accumulation”: 1,
“optimizer”: “adamw_torch”,
“scheduler”: “linear”,
“weight_decay”: 0,
“max_grad_norm”: 1,
“seed”: 42,
“apply_chat_template”: false,
“quantization”: “int4”,
“target_modules”: “”,
“merge_adapter”: true,
“peft”: true,
“lora_r”: 16,
“lora_alpha”: 32,
“lora_dropout”: 0.05
}

----Output logs with error-----

File “/app/env/lib/python3.10/site-packages/accelerate/commands/launch.py”, line 643, in simple_launcher
raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command ‘[’/app/env/bin/python’, ‘-m’, ‘autotrain.trainers.clm’, ‘–training_config’, ‘autotrain-test/training_params.json’]’ died with <Signals.SIGKILL: 9>.

1 Like