Modelpublic

mrcuddle/Typescript-QWen2.5-Coder-3B-Instruct

sourceHugging Faceotherupdated 2y agoView on Hugging Face

2likes14downloads

Model Card

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/> <details><summary>See axolotl config</summary>

axolotl version: 0.6.0

yaml

# axolotl_config.yaml

# Model configuration
base_model: Qwen/Qwen2.5-Coder-3B-Instruct
hub_model_id: mrcuddle/Qwen2.5-Coder-3B-Instruct-TS

# Training parameters
learning_rate: 0.0001  # Adjusted for potential stability improvement
train_batch_size: 4  # Increased for better gradient estimates
eval_batch_size: 4  # Increased for better evaluation stability
num_epochs: 1
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
gradient_accumulation_steps: 2
micro_batch_size: 1


# Distributed training settings
distributed_type: GPU
num_devices: 2  # Adjusted to utilize multiple GPUs if available
total_train_batch_size: 8  # Adjusted to match train_batch_size * num_devices * gradient_accumulation_steps
total_eval_batch_size: 8  # Adjusted to match eval_batch_size * num_devices * gradient_accumulation_steps

# Random seed for reproducibility
seed: 42

datasets:
  - path: mhhmm/typescript-instruct-20k
    type: alpaca
    field_instruction: instruction
    field_output: output
    format: "[INST] {instruction} [/INST]\n{output}"
    no_input_format: "[INST] {instruction} [/INST]"
    roles:
      input: ["USER"]
      output: ["ASSISTANT"]

</details><br>

Qwen2.5-Coder-3B-Instruct-TS

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-3B-Instruct on the mhhmm/typescript-instruct-20k dataset.

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: 0.0001
—trainbatchsize: 1
—evalbatchsize: 4
—seed: 42
—gradientaccumulationsteps: 2
—totaltrainbatch_size: 2
—optimizer: Use adamwhf with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_steps: 100
—num_epochs: 1

Training results

Framework versions

—Transformers 4.47.1
—Pytorch 2.5.1+cu121
—Datasets 3.2.0
—Tokenizers 0.21.0