README.md

---
library_name: transformers
license: apache-2.0
base_model: PrimeIntellect/INTELLECT-1-Instruct
tags:
- axolotl
- generated_from_trainer
datasets:
- neginashz/rationale-llama-chat-dataset
model-index:
- name: star-sft-intellect-instruct-5
  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/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
<details><summary>See axolotl config</summary>

axolotl version: `0.6.0`
```yaml
base_model: PrimeIntellect/INTELLECT-1-Instruct
trust_remote_code: true
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

gpu_memory_limit: 

deepspeed: deepspeed_configs/zero2.json

load_in_8bit: 
load_in_4bit:
strict: false

chat_template: llama3
datasets:
  - path: neginashz/rationale-llama-chat-dataset
    type: chat_template
    chat_template: llama3
    field_messages: messages
    message_field_role: role
    message_field_content: content
    roles:
      system:
        - system
      user:
        - user
      assistant:
        - assistant
    #roles_to_train: ["assistant"]  # default
    # Optional[str]. Which EOS tokens to train on in the conversation. Possible values are:
    # - all: train on all EOS tokens
    # - turn (default): train on the EOS token at the end of each trainable turn
    # - last: train on the last EOS token in the conversation
    #train_on_eos: turn

    
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./star-sft-intellect-5

sequence_len: 4096
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true


wandb_project: star-sft-intellect-instruct-5
wandb_entity: 
wandb_watch:
wandb_name: 
wandb_log_model: 

gradient_checkpointing: true
#gradient_clipping: true
gradient_accumulation_steps: 1
#batch_size: 1
micro_batch_size: 1

num_epochs: 1

optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 0.00002

train_on_inputs: false
group_by_length: false

bf16: true
fp16: false
tf32: false

logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps:
eval_steps: 
save_steps:

evals_per_epoch: 16
saves_per_epoch: 4
eval_max_new_tokens: 128

debug:

weight_decay:
fsdp:
fsdp_config:

hub_model_id: neginashz/star-sft-intellect-instruct-5
hub_strategy: 
early_stopping_patience:

resume_from_checkpoint:
auto_resume_from_checkpoints: true

#special_tokens:
#   pad_token: <|end_of_text|>

```

</details><br>

# star-sft-intellect-instruct-5

This model is a fine-tuned version of [PrimeIntellect/INTELLECT-1-Instruct](https://huggingface.co/PrimeIntellect/INTELLECT-1-Instruct) on the neginashz/rationale-llama-chat-dataset dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3364

## 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: 2e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 4
- total_eval_batch_size: 4
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 6
- num_epochs: 1

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.4105        | 0.0664 | 15   | 0.4274          |
| 0.4759        | 0.1327 | 30   | 0.4348          |
| 0.4704        | 0.1991 | 45   | 0.4255          |
| 0.4612        | 0.2655 | 60   | 0.4167          |
| 0.4765        | 0.3319 | 75   | 0.4030          |
| 0.4022        | 0.3982 | 90   | 0.3932          |
| 0.4234        | 0.4646 | 105  | 0.3856          |
| 0.4008        | 0.5310 | 120  | 0.3736          |
| 0.4066        | 0.5973 | 135  | 0.3649          |
| 0.4007        | 0.6637 | 150  | 0.3568          |
| 0.4059        | 0.7301 | 165  | 0.3491          |
| 0.3622        | 0.7965 | 180  | 0.3429          |
| 0.3655        | 0.8628 | 195  | 0.3388          |
| 0.3655        | 0.9292 | 210  | 0.3368          |
| 0.3868        | 0.9956 | 225  | 0.3364          |


### Framework versions

- Transformers 4.47.1
- Pytorch 2.5.1+cu124
- Datasets 3.1.0
- Tokenizers 0.21.0