See axolotl config

axolotl version: 0.4.1

base_model: NousResearch/Llama-2-7b-hf
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: true
load_in_4bit: false
strict: false

datasets:
  - path: formatted_math_ratio_02_emojianswers_10k.jsonl
    ds_type: json
    type: alpaca
val_set_size: 0.05
dataset_prepared_path: 
output_dir: ./outputs/ppml-formatted

hf_use_auth_token: True
hub_model_id: Ritual-Net/answer-emojis
hub_strategy: all_checkpoints

eval_sample_packing: False


sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: ppml
wandb_entity: ritualnah
wandb_watch: 
wandb_name: emojianswers
wandb_log_model: "checkpoint"

lora_modules_to_save:
  - embed_tokens
  - lm_head


gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 3
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
s2_attention:

warmup_steps: 10
evals_per_epoch: 2
eval_table_size:
eval_max_new_tokens: 128
saves_every_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
special_tokens:
  bos_token: "<s>"
  eos_token: "</s>"
  unk_token: "<unk>"
tokens: # these are delimiters
  - "[INST]"
  - "[/INST]"

answer-emojis

This model is a fine-tuned version of NousResearch/Llama-2-7b-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5239

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.0155	0.0082	1	1.2302
0.5161	0.5031	61	0.5744
0.5398	1.0062	122	0.5379
0.4614	1.4990	183	0.5295
0.4323	2.0021	244	0.5178
0.3823	2.4948	305	0.5239

Framework versions

• PEFT 0.11.1
• Transformers 4.42.3
• Pytorch 2.1.2+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1