---
library_name: transformers
license: cc-by-nc-4.0
base_model: Salesforce/xgen-small-4B-instruct-r
tags:
- axolotl
- generated_from_trainer
datasets:
- hardlyworking/HardlyRPv2
model-index:
- name: HoldMy4B
  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: `0.10.0`
```yaml
base_model: Salesforce/xgen-small-4B-instruct-r

load_in_8bit: false
load_in_4bit: false
strict: false

chat_template: chatml
datasets:
  - path: hardlyworking/HardlyRPv2
    type: chat_template
    split: train
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
val_set_size: 0.1
output_dir: ./outputs/out
dataset_prepared_path: last_run_prepared
shuffle_merged_datasets: true

hub_model_id: hardlyworking/HoldMy4B
hub_strategy: "all_checkpoints"
push_dataset_to_hub:
hf_use_auth_token: true

plugins:
  - axolotl.integrations.liger.LigerPlugin
  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
liger_rope: true
liger_rms_norm: true
liger_layer_norm: true
liger_glu_activation: true
liger_fused_linear_cross_entropy: false
cut_cross_entropy: true

sequence_len: 32768
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

wandb_project: Xgen4B
wandb_entity:
wandb_watch:
wandb_name: Xgen4B
wandb_log_model:

evals_per_epoch: 8
eval_table_size:
eval_max_new_tokens: 128

gradient_accumulation_steps: 2
micro_batch_size: 2
num_epochs: 2
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 1e-5

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

gradient_checkpointing: offload
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
s2_attention:

deepspeed:

warmup_ratio: 0.05
saves_per_epoch: 1
debug:
weight_decay: 0.01
fsdp:
fsdp_config:
special_tokens:
   pad_token:
```

</details><br>

# HoldMy4B

This model is a fine-tuned version of [Salesforce/xgen-small-4B-instruct-r](https://huggingface.co/Salesforce/xgen-small-4B-instruct-r) on the hardlyworking/HardlyRPv2 dataset.
It achieves the following results on the evaluation set:
- Loss: 2.1637

## 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: 1e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- total_eval_batch_size: 4
- optimizer: Use OptimizerNames.ADAMW_BNB 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: 24
- training_steps: 480

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log        | 0     | 0    | 2.6420          |
| 2.0119        | 0.125 | 30   | 2.2105          |
| 1.8963        | 0.25  | 60   | 2.1865          |
| 1.8623        | 0.375 | 90   | 2.1787          |
| 1.8528        | 0.5   | 120  | 2.1746          |
| 1.8784        | 0.625 | 150  | 2.1706          |
| 1.9961        | 0.75  | 180  | 2.1686          |
| 1.8748        | 0.875 | 210  | 2.1672          |
| 2.0385        | 1.0   | 240  | 2.1657          |
| 1.9327        | 1.125 | 270  | 2.1646          |
| 1.8509        | 1.25  | 300  | 2.1645          |
| 1.8279        | 1.375 | 330  | 2.1640          |
| 1.8271        | 1.5   | 360  | 2.1638          |
| 1.8589        | 1.625 | 390  | 2.1637          |
| 1.9824        | 1.75  | 420  | 2.1637          |
| 1.8668        | 1.875 | 450  | 2.1637          |
| 2.0332        | 2.0   | 480  | 2.1637          |


### Framework versions

- Transformers 4.52.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1