1.94 kB
eric-ml-nlp's picture
Add FP32/FP16 precision layout; move FP32 weights under FP32/ and add validated FP16 copy
6dac654 verified