metadata
license: apache-2.0
base_model: mistralai/Mistral-7B-Instruct-v0.3
tags:
- generated_from_trainer
model-index:
- name: outputs/mistral
results: []
See axolotl config
axolotl version: 0.4.0
base_model: mistralai/Mistral-7B-Instruct-v0.3
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer
load_in_8bit: false
load_in_4bit: false
strict: false
chat_template: inst
datasets:
- path: augmxnt/ultra-orca-boros-en-ja-v1
type: sharegpt
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/mistral
sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true
eval_sample_packing: false
use_wandb: true
wandb_project: shisa-v2
wandb_entity: augmxnt
wandb_name: shisa-v1-mistral0.3-7b
gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 3
optimizer: paged_adamw_8bit
lr_scheduler: linear
learning_rate: 8e-6
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
warmup_steps: 100
evals_per_epoch: 2
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed: zero3_bf16.json
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
outputs/mistral
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3791
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: 8e-06
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 3
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.8564 | 0.0045 | 1 | 0.7107 |
0.6131 | 0.5023 | 111 | 0.4259 |
0.6077 | 1.0045 | 222 | 0.3715 |
0.4173 | 1.4932 | 333 | 0.3617 |
0.3812 | 1.9955 | 444 | 0.3468 |
0.2408 | 2.4842 | 555 | 0.3791 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1