Upload TrailRAG cross-encoder model for scifact

CECorrelationEvaluator_validation_results.csv

@@ -1,61 +1,11 @@
 epoch,steps,Pearson_Correlation,Spearman_Correlation
-0,50,nan,nan
-0,-1,nan,nan
-1,50,nan,nan
-1,-1,nan,nan
-2,50,nan,nan
-2,-1,nan,nan
-3,50,nan,nan
-3,-1,nan,nan
-4,50,nan,nan
-4,-1,nan,nan
-5,50,nan,nan
-5,-1,nan,nan
-6,50,nan,nan
-6,-1,nan,nan
-7,50,nan,nan
-7,-1,nan,nan
-8,50,nan,nan
-8,-1,nan,nan
-9,50,nan,nan
-9,-1,nan,nan
-0,50,0.643938028666865,0.6195750638470475
-0,-1,0.6353430223005117,0.61328673153935
-1,50,0.7321815163824663,0.6978031746929375
-1,-1,0.7594597319259389,0.723499905201344
-2,50,0.7940297774325786,0.7448845739658155
-2,-1,0.7999395853103902,0.7564165152979745
-3,50,0.8182861749517223,0.7632045954714167
-3,-1,0.8216628172224243,0.7660747662136247
-4,50,0.8260527721414933,0.7644778893178431
-4,-1,0.8279742708426423,0.7662343764225191
-5,50,0.8292038157129754,0.7659972855296955
-5,-1,0.8259922866915419,0.7675569716971913
-6,50,0.8322338172741923,0.7666086081258973
-6,-1,0.829145976789746,0.7651976848715457
-7,50,0.8348155748910613,0.7688157444973349
-7,-1,0.8352473766100807,0.7676391012221562
-8,50,0.8364154144906919,0.7667186306970768
-8,-1,0.8378338325335756,0.7676956621214246
-9,50,0.8366290995959745,0.768009458891338
-9,-1,0.8367051192077513,0.7684363774597881
-0,50,0.6198869552571915,0.5907845666467741
-0,-1,0.6167852784824035,0.5874369278256336
-1,50,0.7107612375232013,0.6521341589074158
-1,-1,0.7673365017113235,0.6963112017009464
-2,50,0.7908662162706517,0.7080775781876707
-2,-1,0.8110705349801741,0.726221222141359
-3,50,0.8234930400053657,0.7316374783712069
-3,-1,0.8287060042291277,0.7363626441956712
-4,50,0.826754648112667,0.7299496975383057
-4,-1,0.8327556273109046,0.7353923253436677
-5,50,0.8317774706370964,0.735524958855812
-5,-1,0.834508650749443,0.739417868782435
-6,50,0.8378228008043469,0.7381605651380725
-6,-1,0.8378391928599542,0.7355358177398473
-7,50,0.8403673509859279,0.7412250973397235
-7,-1,0.8407120488395872,0.7408644272628397
-8,50,0.8399976524933332,0.7397126099205336
-8,-1,0.8405382158167608,0.739885576430523
-9,50,0.8420007617837248,0.7398421408943822
-9,-1,0.8416671872171954,0.739914274909759
+0,-1,0.5772127119562427,0.5279937929206842
+1,-1,0.7005300429809918,0.6137446634635978
+2,-1,0.7683748424262551,0.6612846272379084
+3,-1,0.7843611275297762,0.6691805537950521
+4,-1,0.789357968372054,0.6754331814865389
+5,-1,0.8100126374378949,0.6964778332002823
+6,-1,0.802975545700353,0.6901123843272047
+7,-1,0.8078541949124098,0.68989685696651
+8,-1,0.8040477492335745,0.6857738263371934
+9,-1,0.8090991990220446,0.6892020341033321

README.md

@@ -21,17 +21,17 @@ model-index:
       type: scifact
     metrics:
     - type: mse
-      value: 0.0880229424591114
+      value: 0.1006970178912303
     - type: mae
-      value: 0.1710951999029706
+      value: 0.1839922902587623
     - type: rmse
-      value: 0.2966866064707193
+      value: 0.3173279343064999
     - type: r2_score
-      value: 0.460256544798364
+      value: 0.4018942599321929
     - type: pearson_correlation
-      value: 0.7952392615312042
+      value: 0.7587210789053855
     - type: spearman_correlation
-      value: 0.7588915635900326
+      value: 0.7092615348799061
 ---
 
 # TrailRAG Cross-Encoder: SciFact Enhanced
@@ -53,20 +53,20 @@ This is a fine-tuned cross-encoder model specifically optimized for **Scientific
 
 | Metric | Value | Description |
 |--------|-------|-------------|
-| **MSE** | **0.088023** | Mean Squared Error (lower is better) |
-| **MAE** | **0.171095** | Mean Absolute Error (lower is better) |
-| **RMSE** | **0.296687** | Root Mean Squared Error (lower is better) |
-| **R² Score** | **0.460257** | Coefficient of determination (higher is better) |
-| **Pearson Correlation** | **0.795239** | Linear correlation (higher is better) |
-| **Spearman Correlation** | **0.758892** | Rank correlation (higher is better) |
+| **MSE** | **0.100697** | Mean Squared Error (lower is better) |
+| **MAE** | **0.183992** | Mean Absolute Error (lower is better) |
+| **RMSE** | **0.317328** | Root Mean Squared Error (lower is better) |
+| **R² Score** | **0.401894** | Coefficient of determination (higher is better) |
+| **Pearson Correlation** | **0.758721** | Linear correlation (higher is better) |
+| **Spearman Correlation** | **0.709262** | Rank correlation (higher is better) |
 
 ### Training Details
 
-- **Training Duration**: 38 minutes
+- **Training Duration**: 32 minutes
 - **Epochs**: 10
 - **Early Stopping**: No
-- **Best Correlation Score**: 0.739914
-- **Final MSE**: 0.088023
+- **Best Correlation Score**: 0.689202
+- **Final MSE**: 0.100697
 
 ### Training Configuration
 

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dfa0fc935b15914786c3f837d2dc6754eb9fb65723bd6da808416cd9c3747c2f
+oid sha256:9695fcb3b98699bb362d2a14f6a7b1960224873016eee1de85fa5239a3660ee0
 size 90866412