See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: peft-internal-testing/tiny-dummy-qwen2
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - e42a4124494711a3_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/e42a4124494711a3_train_data.json
  type:
    field_instruction: question
    field_output: answer
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
device_map: auto
do_eval: true
early_stopping_patience: 2
eval_batch_size: 4
eval_max_new_tokens: 128
eval_steps: 150
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: false
group_by_length: true
hub_model_id: baby-dev/test-09-01
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 50
lora_alpha: 64
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: linear
max_grad_norm: 1.0
max_memory:
  0: 75GB
max_steps: 18000
micro_batch_size: 4
mlflow_experiment_name: /tmp/e42a4124494711a3_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 50
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.95
  adam_epsilon: 1e-5
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 150
saves_per_epoch: null
sequence_len: 1024
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: a23f723a-3c20-47e9-8ceb-7bb7e8892670
wandb_project: SN56-2
wandb_run: your_name
wandb_runid: a23f723a-3c20-47e9-8ceb-7bb7e8892670
warmup_steps: 50
weight_decay: 0.0
xformers_attention: null

test-09-01

This model is a fine-tuned version of peft-internal-testing/tiny-dummy-qwen2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 11.8983

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• training_steps: 6007

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0083	1	11.9304
12.1085	1.2474	150	11.9156
11.917	2.4948	300	11.9073
11.908	3.7422	450	11.9048
11.9016	4.9896	600	11.9026
12.0853	6.2370	750	11.9015
11.9024	7.4844	900	11.9007
11.8993	8.7318	1050	11.9006
11.9043	9.9792	1200	11.9003
12.0835	11.2266	1350	11.9001
11.8991	12.4740	1500	11.9000
11.8963	13.7214	1650	11.8995
11.8964	14.9688	1800	11.8992
12.0746	16.2162	1950	11.8992
11.8988	17.4636	2100	11.8993
11.9032	18.7110	2250	11.8992
11.9002	19.9584	2400	11.8991
12.0821	21.2058	2550	11.8989
11.9004	22.4532	2700	11.8986
11.9018	23.7006	2850	11.8985
11.8981	24.9480	3000	11.8982
12.0775	26.1954	3150	11.8983
11.8959	27.4428	3300	11.8982
11.8987	28.6902	3450	11.8982
11.901	29.9376	3600	11.8982
12.0734	31.1850	3750	11.8983

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1