chatbotNZ/mtext-150224_mistral_merged

See axolotl config

axolotl version: 0.4.0

base_model: mistralai/Mistral-7B-v0.1
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer
is_mistral_derived_model: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: ascherrer/mtext-data-150224_2
    type: completion
    field: text
dataset_prepared_path: last_run_prepared
hub_model_id: ascherrer/mtext-150224_mistral
val_set_size: 0.01
output_dir: ./out

adapter: qlora
lora_model_dir:

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

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


wandb_project: "machine-de-textes"
wandb_entity:
wandb_watch:
wandb_run_id:
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

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_sample_packing: False
eval_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>"
tokens: # these are delimiters
  - "<|s|>"
  - "<|e|>"

mtext-150224_mistral

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0248

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
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• total_eval_batch_size: 4
• 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
4.2902	0.04	1	3.5809
2.6747	0.27	7	2.3159
1.8679	0.53	14	2.0945
1.9268	0.8	21	2.0629
1.6064	1.04	28	2.0900
1.5556	1.3	35	2.0501
1.5276	1.57	42	2.0626
1.517	1.84	49	2.0497
1.4512	2.1	56	2.0396
1.4266	2.36	63	2.0293
1.4217	2.63	70	2.0249
1.4334	2.9	77	2.0248

Framework versions

• PEFT 0.9.1.dev0
• Transformers 4.39.0.dev0
• Pytorch 2.1.2+cu118
• Datasets 2.18.0
• Tokenizers 0.15.0