danube2-1.8b-CodeFeedback

h2o-danube2 with ChatML template

This model was first fine-tuned with BAdam on m-a-p/CodeFeedback-Filtered-Instruction and m-a-p/Code-Feedback, unfiltered from the latest dolphin dataset, using LLama-Factory.

Quants

Thanks to mradermacher!

• mradermacher/danube2-1.8b-CodeFeedback-GGUF

Template

<|im_start|>system
You are a helpful coding assistant.<|im_end|>
<|im_start|>user
{{instruction}}<|im_end|>
<|im_start|>assistant
{{response}}<|im_end|>

BAdam config

System: You are a helpful coding assistant.

### model
model_name_or_path: danube2-base-chatml

### method
stage: sft
do_train: true
finetuning_type: full
use_badam: true
badam_switch_mode: ascending
badam_switch_interval: 50
badam_verbose: 1
badam_start_block: 10
seed: 720

### dataset
dataset: codefeedback_instruct_unfiltered,codefeedback_unfiltered
template: hermes_chatml
cutoff_len: 8192
overwrite_cache: false
preprocessing_num_workers: 12

### output
output_dir: code-feedback-chatml-badam
logging_steps: 5
save_steps: 1
save_strategy: epoch
plot_loss: true
overwrite_output_dir: false

### train
per_device_train_batch_size: 2
gradient_accumulation_steps: 8
learning_rate: 0.00001
num_train_epochs: 1
lr_scheduler_type: cosine
warmup_ratio: 0.01
bf16: true
flash_attn: fa2

### eval
val_size: 0.01
per_device_eval_batch_size: 1
eval_strategy: steps
eval_steps: 2000

BAdam training results