Upload folder using huggingface_hub

README.md

@@ -3,38 +3,35 @@ tags:
 - merge
 - mergekit
 - lazymergekit
-- mlabonne/OrpoLlama-3-8B
+- meta-llama/Meta-Llama-3-8B
 - DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental
 base_model:
-- mlabonne/OrpoLlama-3-8B
+- meta-llama/Meta-Llama-3-8B
 - DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental
 ---
 
 # llama3-discolm-orpo-t2
 
 llama3-discolm-orpo-t2 is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
-* [mlabonne/OrpoLlama-3-8B](https://huggingface.co/mlabonne/OrpoLlama-3-8B)
+* [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B)
 * [DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental](https://huggingface.co/DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental)
 
 ## 🧩 Configuration
 
 ```yaml
-slices:
-  - sources:
-      - model: mlabonne/OrpoLlama-3-8B
-        layer_range: [0, 32]
-      - model: DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental
-        layer_range: [0, 32]
-merge_method: slerp
-base_model: mlabonne/OrpoLlama-3-8B
-parameters:
-  t:
-    - filter: self_attn
-      value: [1, 0.7, 0.5, 0.3, 0.1]
-    - filter: mlp
-      value: [0, 0.3, 0.5, 0.7, 0.9]
-    - value: 0.5
+models:
+  - layer_range: [0, 40]
+    model: meta-llama/Meta-Llama-3-8B
+    parameters:
+      weight: 0.2
+  - layer_range: [0, 40]
+    model: DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental
+    parameters:
+      weight: 0.8
+merge_method: task_arithmetic
+base_model: meta-llama/Meta-Llama-3-8B
 dtype: bfloat16
+random_seed: 0
 ```
 
 ## 💻 Usage