Upload models/hier_binary_web.bin with huggingface_hub

Upload models/hier_binary_tiny.bin with huggingface_hub

Upload models/hier_binary_compact.bin with huggingface_hub

Upload models/hier_binary_quantized.bin with huggingface_hub

Upload models/hier_binary_final.bin with huggingface_hub

Upload models/hier_dialect_web.bin with huggingface_hub

Upload models/hier_dialect_tiny.bin with huggingface_hub

Upload models/hier_dialect_quantized.bin with huggingface_hub

Upload models/hier_dialect_compact.bin with huggingface_hub

Upload models/hier_dialect_final.bin with huggingface_hub

azpieuskalki v2: retrained with SU AZIA Zuberotarra corpus (+6.7K zuberera sentences), 84.06% acc, 12-class

Upload models/azpieuskalki_q.bin with huggingface_hub

Add azpieuskalki bucket=50K quantized (17MB, 82.80%)

Add azpieuskalki full model (233MB, 83.59%)

Add azpieuskalki (sub-dialect) models: 9-class, 83.55% accuracy

Add web dialect model (dim=50)

Add web variant (dim=50, 20K/50K, 34MB total)

Add tiny dialect model

Add tiny binary model

Add compact dialect model

Add compact binary model

Add quantized dialect model

Add quantized binary model

Add hierarchical dialect model (5-class euskalkiak, lr=0.2/150ep, dim=100)

Add hierarchical binary model (batua vs dialectal, lr=3.0/50ep, dim=100)