GitHub - saadtariq001s/custom-INT8-quantizer: Make your own 8-bit linear quantizer that can quantize any LLM, Vision, Audio model to INT8 precision.

saadtariq001s / custom-INT8-quantizer Public

Make your own 8-bit linear quantizer that can quantize any LLM, Vision, Audio model to INT8 precision.

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
L2_linear_I_get_scale_and_zero_point.ipynb		L2_linear_I_get_scale_and_zero_point.ipynb
L2_linear_I_quantize_dequantize_tensor.ipynb		L2_linear_I_quantize_dequantize_tensor.ipynb
L3_linear_II_per_channel.ipynb		L3_linear_II_per_channel.ipynb
L3_linear_II_per_group.ipynb		L3_linear_II_per_group.ipynb
L3_linear_II_per_tensor.ipynb		L3_linear_II_per_tensor.ipynb
L3_linear_II_quantizing_weights_and_activations.ipynb		L3_linear_II_quantizing_weights_and_activations.ipynb
L3_linear_II_symmetric_vs_asymmetric.ipynb		L3_linear_II_symmetric_vs_asymmetric.ipynb
L4_building_quantizer_custom_quantizer.ipynb		L4_building_quantizer_custom_quantizer.ipynb
L4_building_quantizer_load_from_hugging_face_hub.ipynb		L4_building_quantizer_load_from_hugging_face_hub.ipynb
L4_building_quantizer_quantize_models.ipynb		L4_building_quantizer_quantize_models.ipynb
L4_building_quantizer_replace_layers.ipynb		L4_building_quantizer_replace_layers.ipynb
L5_packing_2bit_weights.ipynb		L5_packing_2bit_weights.ipynb
L5_unpacking_2bit_weights.ipynb		L5_unpacking_2bit_weights.ipynb
README.md		README.md
dinner_with_friends.png		dinner_with_friends.png
helper.py		helper.py
helper2.py		helper2.py
helper3.py		helper3.py
requirements.txt		requirements.txt