Upload trained LLM router with ONNX support

README.md

@@ -15,16 +15,16 @@ You can install it with `pip install adaptive-classifier`.
 
 ## Model Details
 
-- Base Model: distilbert/distilbert-base-cased
+- Base Model: benchmark_results/onnx/
 - Number of Classes: 2
-- Total Examples: 616
+- Total Examples: 10
 - Embedding Dimension: 768
 
 ## Class Distribution
 
 ```
-HIGH: 308 examples (50.0%)
-LOW: 308 examples (50.0%)
+HIGH: 5 examples (50.0%)
+LOW: 5 examples (50.0%)
 ```
 
 ## Usage
@@ -33,7 +33,7 @@ LOW: 308 examples (50.0%)
 from adaptive_classifier import AdaptiveClassifier
 
 # Load the model
-classifier = AdaptiveClassifier.from_pretrained("adaptive-classifier/llm-router")
+classifier = AdaptiveClassifier.from_pretrained("adaptive-classifier/model-name")
 
 # Make predictions
 text = "Your text here"
@@ -48,7 +48,7 @@ classifier.add_examples(texts, labels)
 
 ## Training Details
 
-- Training Steps: 20
+- Training Steps: 3688
 - Examples per Class: See distribution above
 - Prototype Memory: Active
 - Neural Adaptation: Active

config.json

@@ -1,8 +1,11 @@
 {
   "config": {
-    "batch_size": 32,
+    "batch_size": 16,
+    "cost_coefficients": {},
+    "cost_function_type": "separable",
     "device_map": "auto",
     "early_stopping_patience": 3,
+    "enable_strategic_mode": false,
     "epochs": 10,
     "ewc_lambda": 100.0,
     "gradient_checkpointing": false,
@@ -17,6 +20,14 @@
     "prototype_weight": 0.8,
     "quantization": null,
     "similarity_threshold": 0.7,
+    "strategic_blend_regular_weight": 0.6,
+    "strategic_blend_strategic_weight": 0.4,
+    "strategic_lambda": 0.1,
+    "strategic_prediction_head_weight": 0.5,
+    "strategic_prediction_proto_weight": 0.5,
+    "strategic_robust_head_weight": 0.2,
+    "strategic_robust_proto_weight": 0.8,
+    "strategic_training_frequency": 10,
     "warmup_steps": 0
   },
   "embedding_dim": 768,
@@ -28,6 +39,10 @@
     "HIGH": 0,
     "LOW": 1
   },
-  "model_name": "distilbert/distilbert-base-cased",
-  "train_steps": 20
+  "model_name": "benchmark_results/onnx/",
+  "train_steps": 3688,
+  "training_history": {
+    "HIGH": 29504,
+    "LOW": 29504
+  }
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:112ac86861bdeed55cb02426a02d16e4c7d9cb94d3bb4df93b1b73bb0c89b42f
+oid sha256:66c6f01605e69e41da15180c100ed0b094a1bba0fcac43e6477acb468a1c7d70
 size 3553504

onnx/config.json

@@ -0,0 +1,24 @@
+{
+  "_name_or_path": "benchmark_results/onnx/",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForMaskedLM"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "initializer_range": 0.02,
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "output_past": true,
+  "pad_token_id": 0,
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "transformers_version": "4.48.0",
+  "vocab_size": 28996
+}

onnx/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d29cdbc87e1fedae5a5f5ff1accab5ec88f10235fb39a4a6e2567e7d5d043bc3
+size 260878383