See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: samoline/5cf4792e-00af-4d5f-b922-4c7d5237f306
bf16: auto
chat_template: llama3
dataloader_num_workers: 8
dataset_prepared_path: null
datasets:
- data_files:
  - 4b98334ce710802c_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/
  type:
    field_input: input
    field_instruction: instruct
    field_output: output
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 4
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 1
gradient_checkpointing: true
group_by_length: false
hub_model_id: cimol/94bee815-bb95-47e6-835f-ad38135a2860
hub_repo: null
hub_strategy: end
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 32
lora_dropout: 0.15
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
loraplus_lr_embedding: 1.0e-06
loraplus_lr_ratio: 16
lr_scheduler: cosine
max_grad_norm: 1
max_steps: 100
micro_batch_size: 8
mlflow_experiment_name: /tmp/4b98334ce710802c_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 20
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.999
  adam_epsilon: 1e-8
optimizer: adamw_torch_fused
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 5
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: 859088c3-5caa-40f5-84fd-490d89cece7b
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 859088c3-5caa-40f5-84fd-490d89cece7b
warmup_steps: 150
weight_decay: 0.0
xformers_attention: null

94bee815-bb95-47e6-835f-ad38135a2860

This model is a fine-tuned version of samoline/5cf4792e-00af-4d5f-b922-4c7d5237f306 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9920

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.999,adam_epsilon=1e-8
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 150
• training_steps: 100

Training results

Training Loss	Epoch	Step	Validation Loss
0.7732	0.0001	1	0.9834
1.1379	0.0002	2	0.9834
0.9396	0.0004	4	0.9834
0.9896	0.0005	6	0.9834
0.8178	0.0007	8	0.9834
0.9894	0.0009	10	0.9834
1.9155	0.0011	12	0.9834
1.0269	0.0012	14	0.9835
0.7556	0.0014	16	0.9836
0.8767	0.0016	18	0.9838
1.1682	0.0018	20	0.9840
0.764	0.0020	22	0.9842
1.1331	0.0021	24	0.9845
0.868	0.0023	26	0.9848
0.7687	0.0025	28	0.9851
0.9999	0.0027	30	0.9855
0.649	0.0029	32	0.9861
1.2312	0.0030	34	0.9865
0.6942	0.0032	36	0.9870
0.8366	0.0034	38	0.9875
0.7883	0.0036	40	0.9879
0.7254	0.0037	42	0.9882
0.9824	0.0039	44	0.9882
0.692	0.0041	46	0.9884
0.7093	0.0043	48	0.9886
0.897	0.0045	50	0.9885
0.549	0.0046	52	0.9886
0.7148	0.0048	54	0.9887
1.327	0.0050	56	0.9888
1.0484	0.0052	58	0.9890
1.1284	0.0053	60	0.9892
0.653	0.0055	62	0.9894
0.8266	0.0057	64	0.9896
0.835	0.0059	66	0.9900
0.7215	0.0061	68	0.9902
0.9994	0.0062	70	0.9904
1.0789	0.0064	72	0.9904
0.7381	0.0066	74	0.9903
0.8917	0.0068	76	0.9903
0.9863	0.0070	78	0.9901
0.8527	0.0071	80	0.9903
0.878	0.0073	82	0.9903
1.2412	0.0075	84	0.9904
1.0756	0.0077	86	0.9904
0.8945	0.0078	88	0.9905
0.6819	0.0080	90	0.9902
0.7676	0.0082	92	0.9901
0.8783	0.0084	94	0.9904
1.0132	0.0086	96	0.9907
0.8469	0.0087	98	0.9913
0.9485	0.0089	100	0.9920

Framework versions

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