Fischerboot/LLama3-Lexi-Aura-3Some-SLERP-SLERP-QLORA

See axolotl config

axolotl version: 0.4.1

base_model: Fischerboot/LLama3-Lexi-Aura-3Some-SLERP-SLERP
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: practical-dreamer/RPGPT_PublicDomain-alpaca
    type: alpaca
dataset_prepared_path: last_run_prepared
val_set_size: 0.1
output_dir: ./outputs/qlora-out

adapter: qlora
lora_model_dir:

sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 4
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:

outputs/qlora-out

This model is a fine-tuned version of Fischerboot/LLama3-Lexi-Aura-3Some-SLERP-SLERP on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0083

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss
1.3184	0.0065	1	1.3193
1.089	0.2545	39	1.1131
1.0475	0.5090	78	1.0719
1.0362	0.7635	117	1.0525
1.0619	1.0049	156	1.0389
1.0165	1.2594	195	1.0322
0.9394	1.5139	234	1.0246
0.999	1.7684	273	1.0182
0.9615	2.0082	312	1.0137
0.9543	2.2626	351	1.0136
0.9429	2.5171	390	1.0109
0.9474	2.7716	429	1.0076
0.8902	3.0098	468	1.0061
0.9144	3.2643	507	1.0089
0.9026	3.5188	546	1.0082
0.9163	3.7732	585	1.0083

Framework versions

• PEFT 0.11.1
• Transformers 4.41.1
• Pytorch 2.1.2+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1