See axolotl config

axolotl version: 0.13.0.dev0

base_model: Qwen/Qwen3-8B
# Automatically upload checkpoint and final model to HF
hub_model_id: okolukisa1/Qwen3-8B-seq-r-1

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: okolukisa1/seq-r-1
    type: alpaca
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/out

sequence_len: 4096
sample_packing: true
eval_sample_packing: true


adapter: qlora
lora_model_dir:
lora_r: 32
lora_alpha: 64
lora_dropout: 0.05
lora_target_linear: true

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 1
optimizer: adamw_torch_fused
lr_scheduler: cosine
learning_rate: 0.0002

bf16: auto
tf32: true

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
resume_from_checkpoint:
logging_steps: 1
flash_attention: true

warmup_ratio: 0.1
evals_per_epoch: 4
saves_per_epoch: 1
weight_decay: 0.0

special_tokens:

# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

Qwen3-8B-seq-r-1

This model is a fine-tuned version of Qwen/Qwen3-8B on the okolukisa1/seq-r-1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4283
• Memory/max Active (gib): 12.88
• Memory/max Allocated (gib): 12.88
• Memory/device Reserved (gib): 17.11

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 62
• training_steps: 623

Training results

Training Loss	Epoch	Step	Validation Loss	Active (gib)	Allocated (gib)	Reserved (gib)
No log	0	0	1.1547	12.2	12.2	12.27
0.4889	0.2497	156	0.4875	12.88	12.88	17.5
0.4674	0.4994	312	0.4510	12.88	12.88	17.11
0.4177	0.7491	468	0.4283	12.88	12.88	17.11

Framework versions

• PEFT 0.18.0
• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.4.1
• Tokenizers 0.22.1