You can deploy this model and start earning money today!

TA-LLM/
ThematicMist-7B-v0.1

Name: ThematicMist-7B-v0.1
Availability: InStock

peft

pytorch

mistral

generated_from_trainer

mistralai/Mistral-7B-Instruct-v0.2

adapter

Views

No views yet

See axolotl config

axolotl version: 0.4.0

yaml

base_model: mistralai/Mistral-7B-Instruct-v0.2
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer
is_mistral_derived_model: true

load_in_8bit: true  # recommended for lora
load_in_4bit: false
strict: false

datasets:
  - path: TA-LLM/alpaca_formatted_data.jsonl  #./data/quote_code.jsonl
    ds_type: json
    type: alpaca
dataset_prepared_path: ./prep #./data
val_set_size: 0.05
output_dir: ./out

sequence_len: 512 #8192
sample_packing: false #true
pad_to_sequence_len: true
eval_sample_packing: false

adapter: lora
lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
  - q_proj
  - v_proj

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 3
micro_batch_size: 2
num_epochs: 2
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.000005

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

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

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_table_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  bos_token: "<s>"
  eos_token: "</s>"
  unk_token: "<unk>"

out

This model was trained from scratch on the TA-LLM/TA_Quote_Code dataset. It achieves the following results on the evaluation set:

Loss: 3.5536

Model description

This is a LoRA fine tuning of mistralai/Mistral-7B-Instruct-v0.2 trained on a novel dataset of quote -> code pairs collected from 30+ thematic analysis projects. The instruction for the data set was:

"Summarize this text into a short code phrase."

Need to cite sources of each study used in dataset:

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: 5e-06
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 3
total_train_batch_size: 6
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
8.0008	0.0	1	13.4803
3.6288	0.25	203	3.8910
2.6522	0.5	406	3.4906
2.696	0.75	609	3.4834
2.896	1.0	812	3.5141
1.9493	1.25	1015	3.5384
2.3697	1.5	1218	3.5356
3.0238	1.75	1421	3.5541
2.1405	2.0	1624	3.5536

Framework versions

PEFT 0.7.1
Transformers 4.37.0
Pytorch 2.0.1
Datasets 2.16.1
Tokenizers 0.15.0

TA-LLM/ThematicMist-7B-v0.1

out