README.md

metadata

license: mit
base_model: BramVanroy/fietje-2b-instruct
tags:
  - trl
  - fietje
  - alignment-handbook
  - dpo
datasets:
  - BramVanroy/ultra_feedback_dutch_cleaned
  - BramVanroy/orca_dpo_pairs_dutch_cleaned
model-index:
  - name: fietje-2b-chat
    results: []
pipeline_tag: text-generation
inference: false
language:
  - nl

Fietje 2B Chat

An open and efficient LLM for Dutch

👱‍♀️ Base version - 🤖 Instruct version - 💬 Chat version (this one) - 🚀 GGUF of chat model

This model is a fine-tuned version of BramVanroy/fietje-2b-sft on the BramVanroy/ultra_feedback_dutch_cleaned and the BramVanroy/orca_dpo_pairs_dutch_cleaned datasets. It achieves the following results on the evaluation set:

• Loss: 0.2842
• Rewards/chosen: -1.1549
• Rewards/rejected: -3.6363
• Rewards/accuracies: 0.8867
• Rewards/margins: 2.4815
• Logps/rejected: -657.6813
• Logps/chosen: -451.3364
• Logits/rejected: -1.2868
• Logits/chosen: -1.3528

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: 2e-06
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-07
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.2515	1.0	1166	0.2842	-1.1549	-3.6363	0.8867	2.4815	-657.6813	-451.3364	-1.2868	-1.3528

Framework versions

• Transformers 4.39.1
• Pytorch 2.1.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2