See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: echarlaix/tiny-random-mistral
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - d1f0ed98551b4cf5_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/d1f0ed98551b4cf5_train_data.json
  type:
    field_instruction: instruction
    field_output: output
    format: '{instruction}'
    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: 4
gradient_checkpointing: true
gradient_clipping: 1.0
group_by_length: false
hub_model_id: ardaspear/b3d14cc8-f59c-47e7-8132-9617849f1baf
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 5.0e-05
load_in_4bit: false
load_in_8bit: false
local_rank: 0
logging_steps: 3
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lr_scheduler: cosine
max_steps: 100
micro_batch_size: 8
mlflow_experiment_name: /tmp/d1f0ed98551b4cf5_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
saves_per_epoch: 4
sequence_len: 1024
special_tokens:
  pad_token: </s>
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: techspear-hub
wandb_mode: online
wandb_name: 5dea7cd1-12a1-49e0-942a-edc220d00ba1
wandb_project: Gradients-On-Five
wandb_run: your_name
wandb_runid: 5dea7cd1-12a1-49e0-942a-edc220d00ba1
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

b3d14cc8-f59c-47e7-8132-9617849f1baf

This model is a fine-tuned version of echarlaix/tiny-random-mistral on the None dataset. It achieves the following results on the evaluation set:

• Loss: 10.3736

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0007	1	10.4094
41.6347	0.0059	9	10.4074
41.6136	0.0117	18	10.4023
41.5916	0.0176	27	10.3970
41.5732	0.0234	36	10.3916
41.5493	0.0293	45	10.3866
41.5311	0.0351	54	10.3819
41.5204	0.0410	63	10.3782
41.5053	0.0468	72	10.3757
41.4921	0.0527	81	10.3743
41.4949	0.0585	90	10.3737
41.4976	0.0644	99	10.3736

Framework versions

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