See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: bigscience/bloomz-560m
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - e687fa7239c356e0_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/e687fa7239c356e0_train_data.json
  type:
    field_instruction: idiom
    field_output: sentence
    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: false
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: dixedus/7fcb8f77-9cc7-4e8b-9235-3bf3bce5b36a
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
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: 200
micro_batch_size: 8
mlflow_experiment_name: /tmp/e687fa7239c356e0_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
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: e42c7ecb-dcd5-460a-9fc7-f330132639a5
wandb_project: Gradients-On-Eight
wandb_run: your_name
wandb_runid: e42c7ecb-dcd5-460a-9fc7-f330132639a5
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

7fcb8f77-9cc7-4e8b-9235-3bf3bce5b36a

This model is a fine-tuned version of bigscience/bloomz-560m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.8318

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
• 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: 200

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0008	1	4.6736
16.9645	0.0129	17	4.1457
15.5833	0.0259	34	3.9829
15.9057	0.0388	51	3.9259
15.7934	0.0518	68	3.8892
16.0657	0.0647	85	3.8695
15.3697	0.0776	102	3.8520
14.9989	0.0906	119	3.8460
15.2332	0.1035	136	3.8393
15.1373	0.1165	153	3.8331
14.9591	0.1294	170	3.8313
15.3457	0.1423	187	3.8318

Framework versions

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