See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: EleutherAI/gpt-neo-1.3B
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 548bfc2cdf0b7cba_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/548bfc2cdf0b7cba_train_data.json
  type:
    field_input: context
    field_instruction: question
    field_output: answer
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: /workspace/axolotl/configs/deepspeed_stage2.json
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: 8
gradient_checkpointing: false
group_by_length: false
hub_model_id: PhoenixB/ac0452cb-1c94-4b2c-a448-663248def925
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 3e-5
liger_fused_linear_cross_entropy: true
liger_glu_activation: true
liger_layer_norm: true
liger_rms_norm: true
liger_rope: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 30
micro_batch_size: 2
mlflow_experiment_name: /tmp/548bfc2cdf0b7cba_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 10
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
plugins:
- axolotl.integrations.liger.LigerPlugin
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 4
sequence_len: 512
special_tokens:
  pad_token: <|endoftext|>
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: e004c32a-fb5a-49bb-9562-e170fb581baf
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: e004c32a-fb5a-49bb-9562-e170fb581baf
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

ac0452cb-1c94-4b2c-a448-663248def925

This model is a fine-tuned version of EleutherAI/gpt-neo-1.3B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2013

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: 3e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• total_eval_batch_size: 16
• 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss
10.5908	0.0036	1	1.3140
9.6421	0.0072	2	1.3143
9.9565	0.0108	3	1.3146
11.0005	0.0144	4	1.3139
10.4604	0.0180	5	1.3140
10.0142	0.0216	6	1.3136
9.7603	0.0252	7	1.3132
10.7314	0.0288	8	1.3119
10.561	0.0324	9	1.3103
10.582	0.0361	10	1.3077
10.4902	0.0397	11	1.3043
10.7695	0.0433	12	1.2995
10.5449	0.0469	13	1.2939
9.9424	0.0505	14	1.2871
10.561	0.0541	15	1.2808
9.9106	0.0577	16	1.2720
9.8418	0.0613	17	1.2628
9.7686	0.0649	18	1.2546
9.7388	0.0685	19	1.2456
10.0977	0.0721	20	1.2390
9.5205	0.0757	21	1.2316
9.1479	0.0793	22	1.2228
9.6904	0.0829	23	1.2168
9.7598	0.0865	24	1.2118
9.5991	0.0901	25	1.2069
10.0308	0.0937	26	1.2047
9.8794	0.0973	27	1.2031
9.5488	0.1009	28	1.2021
9.4438	0.1046	29	1.2014
9.4214	0.1082	30	1.2013

Framework versions

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