Commit History
-
Update model card for the weight-only int8 file (drop stale dynamic-quant metrics)
Verified signaturee3ff41e -
Replace dynamic-range int8 with weight-only int8 (fixes collapsed accuracy)
Verified signatureAs discussed in the collaboration space: the bundled `efficientnet_b1_dynamic_wi8_afp32.tflite` collapses against the torch fp32 reference — Top-1 0/4 with minimum logit correlation 0.32 on fixed random probes (seeds 0-3), 0/10 Top-1 with real photos in the fuller run — which fits EfficientNet-B1's known quantization sensitivity (SE blocks + SiLU activations). The fp32 conversion itself is exact (correlation 1.000000). This PR replaces it with a weight-only int8 quantization of the same model: | variant | Top-1 | Top-5 | min corr | size | |---|---|---|---|---| | dynamic_wi8_afp32 (removed) | 0/4 | 0/20 | 0.320 | 8.5 MB | | weight_only_wi8_afp32 (added) | 4/4 | 18/20 | 0.998 | 8.9 MB | Same ~3.5x size reduction vs fp32. The fp32 model and the Tensor G5 plugin variant are untouched. Happy to share the reproduction harness (torch fp32 reference, fixed-seed probes + photo set) if useful.94ef912
-
Upload efficientnet_b1_Google_Tensor_G5_apply_plugin.tflite with huggingface_hub
Verified signaturee423b81
-
Update optimized LiteRT model files
Verified signaturedce3ba2
-
Updated README.md (#2)
- Updated README.md (d19e7d71f9f258257bd36d144d03e03cd930f942) - Updated README.md (412fc5d7ac6be61799a757346dcc89b7891c497b) Co-authored-by: Akash Verma <akashvverma1995@users.noreply.huggingface.co>161c0da
-
Update README.md (#1)
- updated README.md (801da114a07fbf2cb11bf968942ce4a5ed8fddc4) Co-authored-by: Akash Verma <akashvverma1995@users.noreply.huggingface.co>c55efb9
-
Upload README.md with huggingface_hub
Verified signature43e65d8 -
Upload README.md with huggingface_hub
Verified signature281b780 -
Upload README.md with huggingface_hub
Verified signaturee733f3e -
Initial model upload
Verified signaturec491f60 -
initial commit
Verified signature96c3110