See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/Qwen2.5-Math-1.5B-Instruct
bf16: true
chat_template: llama3
data_processes: 24
dataset_prepared_path: null
datasets:
- data_files:
  - 610f8682142c8104_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/610f8682142c8104_train_data.json
  type:
    field_instruction: instruction
    field_output: response
    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: cimol/dafd1460-9bbc-4fe5-a329-1533e57ce948
hub_repo: null
hub_strategy: end
hub_token: null
learning_rate: 0.0002
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: constant
lr_scheduler_warmup_steps: 50
max_grad_norm: 1.0
max_memory:
  0: 75GB
max_steps: 8972
micro_batch_size: 4
mlflow_experiment_name: /tmp/610f8682142c8104_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 10
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
seed: 17333
sequence_len: 512
strict: false
tf32: true
tokenizer_type: AutoTokenizer
total_train_batch_size: 16
train_batch_size: 4
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 789b7e04-42dd-45e0-b9c3-c954ddcb3301
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 789b7e04-42dd-45e0-b9c3-c954ddcb3301
warmup_steps: 50
weight_decay: 0.0
xformers_attention: null

dafd1460-9bbc-4fe5-a329-1533e57ce948

This model is a fine-tuned version of unsloth/Qwen2.5-Math-1.5B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5346

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: 17333
• 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: constant
• lr_scheduler_warmup_steps: 50
• training_steps: 8972

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0001	1	0.9904
0.6663	0.0088	150	0.6607
0.6431	0.0175	300	0.6296
0.6149	0.0263	450	0.6180
0.594	0.0350	600	0.6102
0.6112	0.0438	750	0.6021
0.5755	0.0526	900	0.5984
0.5957	0.0613	1050	0.5930
0.6007	0.0701	1200	0.5875
0.5867	0.0788	1350	0.5841
0.5836	0.0876	1500	0.5821
0.5919	0.0964	1650	0.5800
0.5829	0.1051	1800	0.5781
0.5651	0.1139	1950	0.5757
0.5653	0.1226	2100	0.5729
0.5699	0.1314	2250	0.5696
0.5485	0.1402	2400	0.5689
0.559	0.1489	2550	0.5680
0.5678	0.1577	2700	0.5656
0.5637	0.1664	2850	0.5647
0.571	0.1752	3000	0.5618
0.5429	0.1840	3150	0.5605
0.5437	0.1927	3300	0.5602
0.5605	0.2015	3450	0.5580
0.5589	0.2102	3600	0.5564
0.5609	0.2190	3750	0.5569
0.5571	0.2278	3900	0.5549
0.5443	0.2365	4050	0.5540
0.5483	0.2453	4200	0.5544
0.5464	0.2541	4350	0.5518
0.5535	0.2628	4500	0.5502
0.5342	0.2716	4650	0.5500
0.5479	0.2803	4800	0.5498
0.5374	0.2891	4950	0.5475
0.5466	0.2979	5100	0.5470
0.5332	0.3066	5250	0.5466
0.5442	0.3154	5400	0.5450
0.5598	0.3241	5550	0.5457
0.5432	0.3329	5700	0.5439
0.5392	0.3417	5850	0.5438
0.5406	0.3504	6000	0.5429
0.537	0.3592	6150	0.5419
0.5545	0.3679	6300	0.5407
0.529	0.3767	6450	0.5408
0.529	0.3855	6600	0.5398
0.5433	0.3942	6750	0.5388
0.5139	0.4030	6900	0.5394
0.542	0.4117	7050	0.5375
0.5408	0.4205	7200	0.5378
0.5401	0.4293	7350	0.5372
0.5312	0.4380	7500	0.5365
0.5409	0.4468	7650	0.5354
0.5366	0.4555	7800	0.5344
0.5338	0.4643	7950	0.5347
0.5211	0.4731	8100	0.5346

Framework versions

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