---
library_name: peft
license: mit
base_model: THUDM/GLM-4-32B-0414
tags:
- axolotl
- generated_from_trainer
datasets:
- anthracite-core/magnum-v5-sft-proto-glm4-instruct-rev1
model-index:
- name: magnum-v5-sft-prototype-glm4-32b-lora
  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: `0.8.0`
```yaml
base_model: THUDM/GLM-4-32B-0414
#base_model_ignore_patterns: "*/*"
# optionally might have model_type or tokenizer_type
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
# Automatically upload checkpoint and final model to HF
hub_model_id: anthracite-core/magnum-v5-sft-prototype-glm4-32b-lora
hub_strategy: "all_checkpoints"
push_dataset_to_hub:
hf_use_auth_token: true


load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: anthracite-core/magnum-v5-sft-proto-glm4-instruct-rev1
    ds_type: parquet
    type:
shuffle_merged_datasets: true
dataset_prepared_path: ./data/magnum-32b-data
val_set_size: 0.01
output_dir: ./data/32b-lora-out

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

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

adapter: lora
lora_model_dir:
lora_r: 128
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
peft_use_rslora: true
lora_modules_to_save:
  - embed_tokens
  - lm_head

wandb_project: 32b-magnum-lora
wandb_entity:
wandb_watch:
wandb_name: run4-Lora-0.001-clip
wandb_log_model:

gradient_accumulation_steps: 2
micro_batch_size: 1
num_epochs: 2
optimizer: paged_ademamix_8bit
lr_scheduler: cosine
learning_rate: 2e-4
max_grad_norm: 0.001

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
s2_attention:

warmup_steps: 40
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens:
saves_per_epoch: 2
debug:
deepspeed: ./deepspeed_configs/zero3_bf16.json
weight_decay: 0.01
fsdp:
fsdp_config:
special_tokens:

```

</details><br>

# magnum-v5-sft-prototype-glm4-32b-lora

This model is a fine-tuned version of [THUDM/GLM-4-32B-0414](https://huggingface.co/THUDM/GLM-4-32B-0414) on the anthracite-core/magnum-v5-sft-proto-glm4-instruct-rev1 dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1075

## 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: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- total_eval_batch_size: 8
- optimizer: Use paged_ademamix_8bit and the args are:
No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 40
- num_epochs: 2.0

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 1.3541        | 0.0024 | 1    | 1.3336          |
| 1.1718        | 0.2503 | 103  | 1.1633          |
| 1.1976        | 0.5006 | 206  | 1.1460          |
| 1.095         | 0.7509 | 309  | 1.1339          |
| 1.1076        | 1.0    | 412  | 1.1213          |
| 1.1063        | 1.2503 | 515  | 1.1128          |
| 1.1214        | 1.5006 | 618  | 1.1089          |
| 1.0286        | 1.7509 | 721  | 1.1075          |


### Framework versions

- PEFT 0.15.1
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.5.0
- Tokenizers 0.21.1