Upload folder using huggingface_hub

README.md

@@ -1,7 +1,7 @@
 ---
 base_model:
-- Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
 - jeffmeloy/Qwen2.5-7B-nerd-uncensored-v1.0
+- Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
 library_name: transformers
 tags:
 - mergekit
@@ -20,8 +20,8 @@ This model was merged using the [SLERP](https://en.wikipedia.org/wiki/Slerp) mer
 ### Models Merged
 
 The following models were included in the merge:
-* [Xiaojian9992024/Qwen2.5-THREADRIPPER-Small](https://huggingface.co/Xiaojian9992024/Qwen2.5-THREADRIPPER-Small)
 * [jeffmeloy/Qwen2.5-7B-nerd-uncensored-v1.0](https://huggingface.co/jeffmeloy/Qwen2.5-7B-nerd-uncensored-v1.0)
+* [Xiaojian9992024/Qwen2.5-THREADRIPPER-Small](https://huggingface.co/Xiaojian9992024/Qwen2.5-THREADRIPPER-Small)
 
 ### Configuration
 
@@ -29,22 +29,22 @@ The following YAML configuration was used to produce this model:
 
 ```yaml
 model_name: "vas-01-7b"
+base_model: Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
+merge_method: slerp
+dtype: bfloat16
 slices:
   - sources:
       - model: Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
         layer_range: [0, 28]
       - model: jeffmeloy/Qwen2.5-7B-nerd-uncensored-v1.0
         layer_range: [0, 28]
-merge_method: slerp
-base_model: Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
 parameters:
   t:
     - filter: self_attn
-      value: [0.5, 0.5, 0.4, 0.4, 0.3]  # More balanced attention weights
+      value: [0.0, 0.3, 0.5, 0.7, 1.0]
     - filter: mlp
-      value: [0.6, 0.6, 0.7, 0.7, 0.8]  # Stronger preservation of Nerd's MLP
+      value: [1.0, 0.7, 0.5, 0.3, 0.0]
     - filter: input_layernorm|post_attention_layernorm
-      value: 0.5  # Equal weight for normalization layers
-    - value: 0.5  # Default for remaining components
-dtype: bfloat16
+      value: 0.5
+    - value: 0.5
 ```

mergekit_config.yml

@@ -1,19 +1,19 @@
 model_name: "vas-01-7b"
+base_model: Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
+merge_method: slerp
+dtype: bfloat16
 slices:
   - sources:
       - model: Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
         layer_range: [0, 28]
       - model: jeffmeloy/Qwen2.5-7B-nerd-uncensored-v1.0
         layer_range: [0, 28]
-merge_method: slerp
-base_model: Xiaojian9992024/Qwen2.5-THREADRIPPER-Small
 parameters:
   t:
     - filter: self_attn
-      value: [0.5, 0.5, 0.4, 0.4, 0.3]  # More balanced attention weights
+      value: [0.0, 0.3, 0.5, 0.7, 1.0]
     - filter: mlp
-      value: [0.6, 0.6, 0.7, 0.7, 0.8]  # Stronger preservation of Nerd's MLP
+      value: [1.0, 0.7, 0.5, 0.3, 0.0]
     - filter: input_layernorm|post_attention_layernorm
-      value: 0.5  # Equal weight for normalization layers
-    - value: 0.5  # Default for remaining components
-dtype: bfloat16
+      value: 0.5
+    - value: 0.5

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95340dfd0059ffcb15a38f7e7b9695c01ea5019da3e9a5c4aca6aaf7470316fb
+oid sha256:b2e9b53831d85a7f95d6687e41863cbc3c49007c28016a1856a3101e4a1abe01
 size 4976698776

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d1432b2eaae5a386a22bf62b91ac6372895f658787a06f5f20f74bcc2a0adc5
+oid sha256:fc4ea59c2afb81ff4936c121053dbca6aec37d6d838d831fc8292414342f6f27
 size 4932751032

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9253eba3b52e42f9ae30b932534df3799b0b1505f758e175c3b9f0a3205f2b6f
+oid sha256:2e63c61767f1571145af318fd332ce38c58c3a433c208bcdb8b9c153bc550cbe
 size 4991495808

model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a4811121f25397ef90999538020985f691e3945ef13f871132f1f69513abf2d
+oid sha256:ca161d09be6665935d5b5f6c6cc55c74180902264efb8800deec008923695cdd
 size 330326240