From 66511d0337f7515dda7ac99116e8ae84e25daa9a Mon Sep 17 00:00:00 2001 From: Dipika Date: Fri, 8 Nov 2024 02:36:18 +0000 Subject: [PATCH] fix group quant bug --- .../compressors/sparse_quantized_compressors/marlin_24.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/compressed_tensors/compressors/sparse_quantized_compressors/marlin_24.py b/src/compressed_tensors/compressors/sparse_quantized_compressors/marlin_24.py index 7954b6b8..4c544588 100644 --- a/src/compressed_tensors/compressors/sparse_quantized_compressors/marlin_24.py +++ b/src/compressed_tensors/compressors/sparse_quantized_compressors/marlin_24.py @@ -238,7 +238,7 @@ def pack_scales_24(scales, quantization_args, w_shape): _, scale_perm_2_4, scale_perm_single_2_4 = get_permutations_24(num_bits) if ( - quantization_args.strategy is QuantizationStrategy.GROUP + quantization_args.strategy == QuantizationStrategy.GROUP and quantization_args.group_size < size_k ): scales = scales.reshape((-1, len(scale_perm_2_4)))[:, scale_perm_2_4]