+ QV-jRt^RIHt^RIHtHt]!4'd ^RIt^RIHt]P!] 4t RRRRR/RRRRR /R RRRR /R RRRR /R RRRR/RRRRR /RRRRR/RRRRR //t Rt RRRlltR#)z;AWQ (Activation aware Weight Quantization) integration file)should_convert_module)is_torch_availableloggingN starcoder2actlayer_before_actc_fcRefinedWebModel dense_h_to_4hfalconmptup_projgptjfc_ingpt_neox gpt_bigcodebloom gelu_implc^RIHpV\9dV#VP4Fwr4\V,R,p\V,R,pW58Xdg\ W4'dV\ V\V,R,4pVP p\P!V4p V!WI4VPV&\WA4p K V#))ScaledActivationrr) &gptqmodel.quantization.awq.modules.actrAWQ_SCALES_MAPPINGSnamed_childrenhasattrgetattr out_featurestorchones_modulesreplace_quantization_scales) model model_typernamemoduleact_namelayer_before_act_namersize scale_like_s && n/Volumes/fast/ai/experiments/ui-tars-smoke/.venv/lib/python3.14/site-packages/transformers/integrations/awq.pyr r 'sG,, ,,. &z259 3J ?@R S   E E&u.A*.MN`.ab #00DD)J#3F#GENN4 ' ;/ LcTV^8dQhR\\,R,R\/#) device_mapNreturn)strdictbool)formats"r* __annotate__r48s)??d T! ?  ?r+c&^RIHp^RIHpV!VPVP RRVP VPVVPVPRR7 pVP4Fwrx\Wq4'gK\P!R4;_uu_4\V\P 4'dV!VPVP"VP$VP VP&VP(VP*RJVP,PRR7 p V P/R4VP1Wy4Rp RRR4K X 'g\2P5R 4V# +'giEK%;i) a Public method that replaces the linear layers of the given model with awq quantized layers. Args: model (`torch.nn.Module`): The model to convert, can be any `torch.nn.Module` instance. quantization_config (`AwqConfig`): The quantization config object that contains the quantization parameters. modules_to_not_convert (`list[str]`, *optional*, defaults to `None`): A list of nn.Linear weights to not convert. If a parameter path is in the list (e.g. `lm_head.weight`), the corresponding module will not be converted. device_map (`Union[str, dict]`, *optional*, defaults to `None`): The device map that maps the parameters to the device )METHOD)hf_select_quant_linear_v2F) bits group_sizedesc_actsymr3backendr. quant_method zero_pointpackmetaNT) r8r;r:r9 in_featuresrbiasdevregister_bufferszYou are loading your model using eetq but no linear modules were found in your model. Please double check your model architecture, or submit an issue on github if you think this is a bug.)gptqmodel.quantizationr6gptqmodel.utils.importerr7r8r9r3r<AWQr> named_modulesrrdevice isinstancennLinearr;r:rArrBweightrequires_grad_ set_submoduleloggerwarning) r!modules_to_not_convertquantization_configr.r6r7 target_cls module_namer$ new_modulehas_been_replaceds &&&& r*replace_with_awq_linearrX8sK(.B* % %&11 "))#++ZZ&11  J %224 $[II  \\& ! !&")),,',11+//0992== & 2 2!'!4!4D0 ,,%)  ))%0##K<$(!" ! 5(   L1" ! !s !B2E>> F )NNN)__doc__quantizers.quantizers_utilsrutilsrrrtorch.nnrK get_logger__name__rPrr rXr+r*r`s>?/   H %5%!3V<u&8/J ue/A E5,i 8 UE-w 71?CE5"4f= e["4o F "??r+