See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/Qwen2-0.5B-Instruct
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 297768cf1b07a85c_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/297768cf1b07a85c_train_data.json
  type:
    field_instruction: database
    field_output: text
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 4
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 6
gradient_checkpointing: true
group_by_length: false
hub_model_id: Romain-XV/1ca5001e-174c-4093-bb47-1ebd25d8da14
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_best_model_at_end: true
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.3
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lora_target_modules:
- q_proj
- k_proj
- v_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 5376
micro_batch_size: 4
mlflow_experiment_name: /tmp/297768cf1b07a85c_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
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: 100
sequence_len: 1024
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.04688584235104368
wandb_entity: null
wandb_mode: online
wandb_name: d72718f4-c991-42fc-ada7-d52f3778f0fd
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: d72718f4-c991-42fc-ada7-d52f3778f0fd
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

1ca5001e-174c-4093-bb47-1ebd25d8da14

This model is a fine-tuned version of unsloth/Qwen2-0.5B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3620

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 6
• total_train_batch_size: 24
• optimizer: Use OptimizerNames.ADAMW_BNB 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: 10
• training_steps: 5376

Training results

Training Loss	Epoch	Step	Validation Loss
3.16	0.0002	1	3.2104
1.6527	0.0236	100	1.7146
1.5718	0.0472	200	1.4255
1.0595	0.0708	300	1.2537
1.1479	0.0944	400	1.1307
1.0057	0.1181	500	1.0392
1.091	0.1417	600	0.9590
1.0414	0.1653	700	0.9006
0.7725	0.1889	800	0.8477
0.8099	0.2125	900	0.8059
0.8009	0.2361	1000	0.7638
0.6346	0.2597	1100	0.7316
0.9218	0.2833	1200	0.7010
0.5901	0.3070	1300	0.6727
0.5379	0.3306	1400	0.6462
1.0677	0.3542	1500	0.6253
0.6949	0.3778	1600	0.6083
0.5041	0.4014	1700	0.5854
0.79	0.4250	1800	0.5689
0.5347	0.4486	1900	0.5519
0.6645	0.4722	2000	0.5365
0.6224	0.4958	2100	0.5251
0.6407	0.5195	2200	0.5119
1.0207	0.5431	2300	0.5016
0.6107	0.5667	2400	0.4900
0.4412	0.5903	2500	0.4785
0.4818	0.6139	2600	0.4681
0.6567	0.6375	2700	0.4577
0.3419	0.6611	2800	0.4503
0.2586	0.6847	2900	0.4405
0.6874	0.7084	3000	0.4336
0.4483	0.7320	3100	0.4262
0.5098	0.7556	3200	0.4184
0.3629	0.7792	3300	0.4113
0.5132	0.8028	3400	0.4058
0.6217	0.8264	3500	0.4011
0.7283	0.8500	3600	0.3956
0.3335	0.8736	3700	0.3904
0.445	0.8972	3800	0.3867
0.4191	0.9209	3900	0.3823
0.6112	0.9445	4000	0.3784
0.3837	0.9681	4100	0.3752
0.5076	0.9917	4200	0.3722
0.3469	1.0153	4300	0.3702
0.5091	1.0389	4400	0.3684
0.2305	1.0625	4500	0.3674
0.4263	1.0861	4600	0.3654
0.5123	1.1098	4700	0.3644
0.5012	1.1334	4800	0.3637
0.3304	1.1570	4900	0.3629
0.3847	1.1806	5000	0.3624
0.2313	1.2042	5100	0.3621
0.4293	1.2278	5200	0.3620
0.3886	1.2514	5300	0.3620

Framework versions

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