AbstractPhil commited on
Commit
c285ee0
Β·
verified Β·
1 Parent(s): cbf8cc2

Update README - Run 20251012_151647

Browse files
Files changed (1) hide show
  1. README.md +16 -16
README.md CHANGED
@@ -21,7 +21,7 @@ model-index:
21
  type: imagenet-1k
22
  metrics:
23
  - type: accuracy
24
- value: 76.86
25
  ---
26
 
27
  # David: Multi-Scale Crystal Classifier
@@ -51,15 +51,15 @@ as class prototypes with role-weighted similarity computation (Rose Loss).
51
  ## Performance
52
 
53
  ### Best Results
54
- - **Validation Accuracy**: 76.86%
55
- - **Best Epoch**: 9
56
- - **Final Train Accuracy**: 91.25%
57
 
58
  ### Per-Scale Performance
59
- - **Scale 256**: 70.23%
60
- - **Scale 512**: 74.51%
61
- - **Scale 768**: 75.80%
62
- - **Scale 1024**: 75.58%
63
 
64
 
65
  ## Usage
@@ -77,18 +77,18 @@ AbstractPhil/gated-david/
77
  β”œβ”€β”€ best_model.json # Latest best model info
78
  β”œβ”€β”€ weights/
79
  β”‚ └── david_hierarchical/
80
- β”‚ └── 20251012_145649/
81
  β”‚ β”œβ”€β”€ MODEL_SUMMARY.txt # 🎯 Human-readable performance summary
82
  β”‚ β”œβ”€β”€ training_history.json # πŸ“ˆ Epoch-by-epoch training curve
83
- β”‚ β”œβ”€β”€ best_model_acc76.86.safetensors # ⭐ Accuracy in filename!
84
- β”‚ β”œβ”€β”€ best_model_acc76.86_metadata.json
85
  β”‚ β”œβ”€β”€ final_model.safetensors
86
  β”‚ β”œβ”€β”€ checkpoint_epoch_X_accYY.YY.safetensors
87
  β”‚ β”œβ”€β”€ david_config.json
88
  β”‚ └── train_config.json
89
  └── runs/
90
  └── david_hierarchical/
91
- └── 20251012_145649/
92
  └── events.out.tfevents.* # TensorBoard logs
93
  ```
94
 
@@ -102,8 +102,8 @@ from huggingface_hub import hf_hub_download
102
 
103
  # Specify model variant and run
104
  model_name = "david_hierarchical"
105
- run_id = "20251012_145649"
106
- accuracy = "76.86" # From MODELS_INDEX.json
107
 
108
  # Download config
109
  config_path = hf_hub_download(
@@ -194,7 +194,7 @@ score = w_anchor * sim(z, anchor) + w_need * sim(z, need) + ...
194
  author = {AbstractPhil},
195
  year = {2025},
196
  url = {https://huggingface.co/AbstractPhil/gated-david},
197
- note = {Run ID: 20251012_145649}
198
  }
199
  ```
200
 
@@ -209,4 +209,4 @@ Special thanks to Claude (Anthropic) for debugging assistance.
209
 
210
  ---
211
 
212
- *Generated on 2025-10-12 15:15:01*
 
21
  type: imagenet-1k
22
  metrics:
23
  - type: accuracy
24
+ value: 72.23
25
  ---
26
 
27
  # David: Multi-Scale Crystal Classifier
 
51
  ## Performance
52
 
53
  ### Best Results
54
+ - **Validation Accuracy**: 72.23%
55
+ - **Best Epoch**: 0
56
+ - **Final Train Accuracy**: 68.42%
57
 
58
  ### Per-Scale Performance
59
+ - **Scale 256**: 65.87%
60
+ - **Scale 512**: 69.96%
61
+ - **Scale 768**: 71.41%
62
+ - **Scale 1024**: 71.52%
63
 
64
 
65
  ## Usage
 
77
  β”œβ”€β”€ best_model.json # Latest best model info
78
  β”œβ”€β”€ weights/
79
  β”‚ └── david_hierarchical/
80
+ β”‚ └── 20251012_151647/
81
  β”‚ β”œβ”€β”€ MODEL_SUMMARY.txt # 🎯 Human-readable performance summary
82
  β”‚ β”œβ”€β”€ training_history.json # πŸ“ˆ Epoch-by-epoch training curve
83
+ β”‚ β”œβ”€β”€ best_model_acc72.23.safetensors # ⭐ Accuracy in filename!
84
+ β”‚ β”œβ”€β”€ best_model_acc72.23_metadata.json
85
  β”‚ β”œβ”€β”€ final_model.safetensors
86
  β”‚ β”œβ”€β”€ checkpoint_epoch_X_accYY.YY.safetensors
87
  β”‚ β”œβ”€β”€ david_config.json
88
  β”‚ └── train_config.json
89
  └── runs/
90
  └── david_hierarchical/
91
+ └── 20251012_151647/
92
  └── events.out.tfevents.* # TensorBoard logs
93
  ```
94
 
 
102
 
103
  # Specify model variant and run
104
  model_name = "david_hierarchical"
105
+ run_id = "20251012_151647"
106
+ accuracy = "72.23" # From MODELS_INDEX.json
107
 
108
  # Download config
109
  config_path = hf_hub_download(
 
194
  author = {AbstractPhil},
195
  year = {2025},
196
  url = {https://huggingface.co/AbstractPhil/gated-david},
197
+ note = {Run ID: 20251012_151647}
198
  }
199
  ```
200
 
 
209
 
210
  ---
211
 
212
+ *Generated on 2025-10-12 15:18:32*