See axolotl config

axolotl version: 0.4.0

base_model: premai-io/prem-1B
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: argilla/distilabel-capybara-dpo-7k-binarized
    type: orpo.chat_template
dataset_prepared_path: last_run_prepared
val_set_size: 0.001
output_dir: ./prem-1B-32k
save_safetensors: true
sequence_len: 8192
sample_packing: false
pad_to_sequence_len: false
use_pose: true
pose_max_context_len: 262144
min_sample_len: 6144
pose_num_chunks: 16
curriculum_sampling: true

overrides_of_model_config:
  rope_theta: 500000.0
  max_position_embeddings: 262144

  # peft_use_dora: true
adapter: lora
peft_use_rslora: true
lora_model_dir:
lora_r: 1024
lora_alpha: 1024
lora_dropout: 0.1
lora_target_modules:
  - q_proj
  - k_proj
  - v_proj
  - o_proj
lora_modules_to_save:
  - embed_tokens
  - lm_head

wandb_project:
wandb_entity: 
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 8
micro_batch_size: 1
num_epochs: 20
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.00001
max_grad_norm: 1.0
adam_beta2: 0.95

train_on_inputs: false
group_by_length: false
bf16: true
fp16:
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
sdp_attention:
s2_attention:

warmup_steps: 10
evals_per_epoch: 8
saves_per_epoch: 8
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  pad_token: <|end_of_text|>

prem-1B-32k

This model is a fine-tuned version of premai-io/prem-1B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0059

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
0.7672	1.0	1	3.0074
0.7672	2.0	2	2.6057
0.7422	3.0	3	2.2898
0.7211	4.0	4	2.1453
0.6591	5.0	5	1.6360
0.4514	6.0	6	0.7589
0.24	7.0	7	0.6621
0.1584	8.0	8	0.8121
0.1235	9.0	9	0.7538
0.0998	10.0	10	0.7743
0.0869	11.0	11	0.7771
0.1692	12.0	12	0.8293
0.0702	13.0	13	0.8939
0.063	14.0	14	0.9582
0.0567	15.0	15	0.9825
0.052	16.0	16	0.9960
0.0488	17.0	17	0.9883
0.0457	18.0	18	1.0004
0.0436	19.0	19	1.0056
0.0427	20.0	20	1.0059

Framework versions

• PEFT 0.10.0
• Transformers 4.40.0.dev0
• Pytorch 2.1.2+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0