Safetensors
qwen2
spectrum
sft
dpo
Eval Results
DavidGF commited on
Commit
6db07c0
1 Parent(s): b828644

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -2
README.md CHANGED
@@ -76,8 +76,8 @@ This model extends our two-phase SFT model with an additional DPO phase, creatin
76
 
77
  **Dataset Composition for DPO**:
78
  - Extended previous DPO dataset
79
- - New SauerkrautLM-Fermented-GER-DPO dataset
80
- - SauerkrautLM-Fermented-Irrelevance-GER-DPO dataset
81
  - Carefully balanced to maintain German language capabilities
82
 
83
  ## Released Datasets
@@ -105,6 +105,7 @@ This DPO-enhanced version aims to:
105
  - Provide valuable training resources to the community
106
 
107
  ## Evaluation
 
108
 
109
  **AGIEVAL**
110
  ![SauerkrautLM-v2-14b-DPO-AGIEVAL](https://vago-solutions.ai/wp-content/uploads/2024/11/SauerkrautLM-v2-14b-DPO-AGIEVAL.png "SauerkrautLM-v2-14b-DPO-AGIEVAL")
 
76
 
77
  **Dataset Composition for DPO**:
78
  - Extended previous DPO dataset
79
+ - New SauerkrautLM-Fermented-GER-DPO dataset (release soon)
80
+ - SauerkrautLM-Fermented-Irrelevance-GER-DPO dataset (release soon)
81
  - Carefully balanced to maintain German language capabilities
82
 
83
  ## Released Datasets
 
105
  - Provide valuable training resources to the community
106
 
107
  ## Evaluation
108
+ (same diagrams as in SauerkrautLM-v2-14b-SFT model card)
109
 
110
  **AGIEVAL**
111
  ![SauerkrautLM-v2-14b-DPO-AGIEVAL](https://vago-solutions.ai/wp-content/uploads/2024/11/SauerkrautLM-v2-14b-DPO-AGIEVAL.png "SauerkrautLM-v2-14b-DPO-AGIEVAL")