Dessa

Results 9 comments of Dessa

NUC8i7BEH?

I have the same issue with bean canyon nucs. Posted question on Intel NUC forums. Hades WMI commands are not supported either. >>>> Doing some test on NUC8i7BE with the...

experimental.linear_quantize_activations with Classifier_config in ct.conver "fails" w/message 'dict' object has no attribute 'flatten'

I'm getting bunch of Running compression pass linear_quantize_activations: calibrating sample 3[/20](http://localhost:8888/20) fails. in Jupyter lab yes, it is 8.0b1 , experimental linear_quantize_activations is only available on 8.0b1.... macOS 15, pip...

coremltools 8.0b1 linear_quantize_weights is not working for models converted with minimum_deployment_target=ct.target.iOS18

Thank you , could you confirm int4 works as well ? iOS8 target is required for sure in this case. This was the main motivation for the test, and it...

coremltools 8.0b1 linear_quantize_weights is not working for models converted with minimum_deployment_target=ct.target.iOS18

Indeed, size is 1/2 for 4-bit lin quant, and MD5 for weight.bin is the same for iOS17 and iOS18 export for 8 bt. Exporting model's Specs I can see INT4...

coremltools 8.0b1 linear_quantize_weights is not working for models converted with minimum_deployment_target=ct.target.iOS18

FB14257215 Xcode beta 3 has the same issue

coremltools 8.0b1 linear_quantize_weights is not working for models converted with minimum_deployment_target=ct.target.iOS18

This issue is resolved in Xcode beta 4, both Int4 and int8 are showing up correctly Version 16.0 beta 4 (16A5211f) FB14257215

Model state doesn't work with transpose

Confirmed, IMHO it is H/W limitation due to the same ANE block is used for both slice_update and transpose/permute? Adding "non-optimizable op" like add external tensor helps, depending on tensor...

Model state doesn't work with transpose

Xcode profile Pic, that its on ANE

[Quantization] int8 compute with PostTraining Quantization (W8A8)

Could be this, but due to "1" for channel in conv https://machinelearning.apple.com/research/neural-engine-transformers For example, if the last axis is used as a singleton one by the model implementation’s data format,...