YoungPanda/test4

See axolotl config

axolotl version: 0.4.0

base_model: upstage/SOLAR-10.7B-v1.0
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer
is_llama_derived_model: true

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: Drewskidang/rag
    type: alpaca
  - path: Drewskidang/chatinstruct
    type: alpaca
  - path: Drewskidang/Instruct
    type: alpaca
  - path: Drewskidang/output
    type: alpaca
dataset_prepared_path: last_run_prepared
val_set_size: 0.05
output_dir: ./out

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true

adapter:
lora_model_dir:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:
lora_fan_in_fan_out:

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 10
num_epochs: 2
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
flash_attn_cross_entropy: false
flash_attn_rms_norm: true
flash_attn_fuse_qkv: false
flash_attn_fuse_mlp: true

warmup_steps: 100
evals_per_epoch: 4
eval_table_size:
saves_per_epoch: 1
debug:
deepspeed: deepspeed_configs/zero2.json # multi-gpu only
weight_decay: 0.1
fsdp:
fsdp_config:
special_tokens:

out

This model is a fine-tuned version of upstage/SOLAR-10.7B-v1.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 12.2158

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: 10
• eval_batch_size: 10
• seed: 42
• distributed_type: multi-GPU
• num_devices: 7
• total_train_batch_size: 70
• total_eval_batch_size: 70
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.0045	0.04	1	1.0213
0.9752	0.26	7	0.9097
0.8562	0.52	14	0.8622
0.8457	0.78	21	0.8343
0.8846	1.04	28	0.8288
0.596	1.26	35	0.9086
1.4506	1.52	42	2.4507
27.7592	1.78	49	12.2158

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0