---
library_name: peft
license: llama3.2
base_model: meta-llama/Llama-3.2-1B
tags:
- generated_from_trainer
datasets:
- teknium/OpenHermes-2.5
model-index:
- name: outputs/qlora-out
  results: []
---


"Don't learn by adding vocabulary to vocab."

This model is completely broken. I figured it out after 6 hours of training.

```log
Traceback (most recent call last):
 File "/data/minpeter/qlora-llama-1b-chatml-v2/merge.py", line 6, in <module>
 model = PeftModel.from_pretrained(base_model, peft_model_id)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/home/minpeter/anaconda3/envs/axo/lib/python3.12/site-packages/peft/peft_model.py", line 581, in from_pretrained
 load_result = model.load_adapter(
 ^^^^^^^^^^^^^^^^^^^^
 File "/home/minpeter/anaconda3/envs/axo/lib/python3.12/site-packages/peft/peft_model.py", line 1239, in load_adapter
 load_result = set_peft_model_state_dict(
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/home/minpeter/anaconda3/envs/axo/lib/python3.12/site-packages/peft/utils/save_and_load.py", line 451, in set_peft_model_state_dict
 load_result = model.load_state_dict(peft_model_state_dict, strict=False)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/home/minpeter/anaconda3/envs/axo/lib/python3.12/site-packages/torch/nn/modules/module.py", line 2584, in load_state_dict
 raise RuntimeError(
RuntimeError: Error(s) in loading state_dict for PeftModelForCausalLM:
 size mismatch for base_model.model.model.embed_tokens.modules_to_save.default.weight: copying a param with shape torch.Size([128258, 2048]) from checkpoint, the shape in current model is torch.Size([128256, 2048]).
 size mismatch for base_model.model.lm_head.modules_to_save.default.weight: copying a param with shape torch.Size([128258, 2048]) from checkpoint, the shape in current model is torch.Size([128256, 2048]).
```

<!-- 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: meta-llama/Llama-3.2-1B

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: teknium/OpenHermes-2.5
    type: chat_template
    chat_template: chatml
    field_messages: conversations
    message_field_role: from
    message_field_content: value
    shards: 1

save_safetensors: true
auto_resume_from_checkpoints: true
save_steps: 200

chat_template: chatml
dataset_prepared_path: last_run_prepared
val_set_size: 0.1
output_dir: ./outputs/qlora-out

adapter: qlora
lora_model_dir:

sequence_len: 2048
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project: "axolotl"
wandb_entity: "kasfiekfs-e"
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 1
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_max_new_tokens: 128
# saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:

tokens:
  - <|im_start|>

special_tokens:
  eos_token: <|im_end|>
  pad_token: <|end_of_text|>

lora_modules_to_save:
  - lm_head
  - embed_tokens

# <--- unsloth config --->
unsloth_lora_mlp: true
unsloth_lora_qkv: true
unsloth_lora_o: true
```

</details><br>

# outputs/qlora-out

This model is a fine-tuned version of [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B) on the teknium/OpenHermes-2.5 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8874

## 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
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- 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: 10
- num_epochs: 1.0

### Training results

| Training Loss | Epoch  | Step  | Validation Loss |
|:-------------:|:------:|:-----:|:---------------:|
| 1.2931        | 0.0000 | 1     | 1.3097          |
| 1.0603        | 0.2500 | 5419  | 0.9798          |
| 0.7964        | 0.5000 | 10838 | 0.9218          |
| 0.7602        | 0.7500 | 16257 | 0.8874          |


### Framework versions

- PEFT 0.14.0
- Transformers 4.47.1
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0