---
license: mit
base_model: microsoft/Phi-3-mini-128k-instruct
tags:
- generated_from_trainer
model-index:
- name: phi3-sft-out
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
<details><summary>See axolotl config</summary>

axolotl version: `0.4.0`
```yaml
base_model: microsoft/Phi-3-mini-128k-instruct
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
trust_remote_code: true

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: sosoai/mixed_dataset
    type: alpaca

dataset_prepared_path:
val_set_size: 0.05
output_dir: ./phi3-sft-out

sequence_len: 2048
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: 2
num_epochs: 5
optimizer: adamw_torch
adam_beta2: 0.95
adam_epsilon: 0.00001
max_grad_norm: 1.0
lr_scheduler: cosine
learning_rate: 0.000003

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

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: True
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 100
eval_sample_packing: False
evals_per_epoch: 4
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.1
fsdp:
fsdp_config:
resize_token_embeddings_to_32x: true
special_tokens:
  pad_token: "<|endoftext|>"

```

</details><br>

# phi3-sft-out

This model is a fine-tuned version of [microsoft/Phi-3-mini-128k-instruct](https://huggingface.co/microsoft/Phi-3-mini-128k-instruct) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2406

## 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: 3e-06
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-05
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 4.6772        | 0.0   | 1    | 1.3794          |
| 3.1471        | 0.25  | 175  | 1.2942          |
| 3.0306        | 0.5   | 350  | 1.2572          |
| 2.7486        | 0.75  | 525  | 1.2491          |
| 2.7702        | 1.0   | 700  | 1.2467          |
| 2.6302        | 1.24  | 875  | 1.2458          |
| 2.8356        | 1.49  | 1050 | 1.2436          |
| 2.7697        | 1.74  | 1225 | 1.2418          |
| 2.7226        | 2.0   | 1400 | 1.2415          |
| 2.7363        | 2.23  | 1575 | 1.2411          |
| 2.6754        | 2.48  | 1750 | 1.2407          |
| 2.9697        | 2.73  | 1925 | 1.2407          |
| 2.6213        | 2.99  | 2100 | 1.2406          |
| 2.6752        | 3.23  | 2275 | 1.2407          |
| 2.7226        | 3.48  | 2450 | 1.2404          |
| 2.6131        | 3.73  | 2625 | 1.2405          |
| 2.7255        | 3.98  | 2800 | 1.2404          |
| 2.7335        | 4.21  | 2975 | 1.2404          |
| 2.7924        | 4.46  | 3150 | 1.2406          |
| 2.6851        | 4.71  | 3325 | 1.2406          |


### Framework versions

- Transformers 4.40.0.dev0
- Pytorch 2.1.1
- Datasets 2.15.0
- Tokenizers 0.15.0