+ TV-j-^RIt^RIHt^RIHt^RIHtHt^RIH t ^RI H t ^RI HtHtHt^RIHtHtHt^RIHtHt^RIHt^R IHt]!R R ]44tRR lt]!] P<R R7RRRll4t!RR] P@4t!!RR] P@4t"!RR] P@4t#!RR] P@4t$!RR] P@4t%!RR] P@4t&R#)N) dataclass)partial)AnyOptional) shard_inplace shard_linear sum_gradients) BaseModelArgscreate_attention_maskscaled_dot_product_attention)KVCacheRotatingKVCache)initialize_rope) SwitchGLUcha]tRt^toRt^$t^t^tRtRt Rt Rt ^@t ^@t ^t^tRtRtRtV3RltRtVtR#) ModelArgsgpt_ossigh㈵>i@ iINc<V^8dQh/S[;R&S[;R&S[;R&S[;R&S[;R&S[;R&S[;R&S[;R&S[;R &S[;R &S[;R &S[;R &S[;R &S[;R&S[;R&#) model_typenum_hidden_layersnum_local_expertsnum_experts_per_tok vocab_size rms_norm_eps hidden_sizeintermediate_sizehead_dimnum_attention_headsnum_key_value_headssliding_window rope_theta rope_scaling layer_types)strintfloatrlist)format __classdict__s"f/Volumes/fast/ai/experiments/ui-tars-smoke/.venv/lib/python3.14/site-packages/mlx_lm/models/gpt_oss.py __annotate__ModelArgs.__annotate__s     !!  !)__name__ __module__ __qualname____firstlineno__rrrrrrrrrrr r!r"r#r$__annotate_func____static_attributes____classdictcell__)r*s@r+rrs]J  JLK!H! NJLK!r.rc\P!W)VR7pVRV)R13,p\P!WVR7pWT3#)zMLX equivalent of torch.topk)kthaxis.Nr9)mx argpartitiontake_along_axis)akr9partitioned_indices top_k_indices top_k_valuess&&& r+mlx_topkrC&sB//!$?'aRS1M%%aTBL  &&r.T) shapelessc0V^8dQhR\R\/#)ralphalimit)r')r)s"r+r,r,1s $ $5 $ $r.c\P!VRVR7p\P!W)VR7pW!,p\P!V4pW,pW`^,,#)N)a_mina_max)r;clipsigmoid)x_linearx_glurFrG glu_scaledsigout_glus&&&& r+swiglurR0sQ GGEU 3EwwxvU;HJ **Z CkG l ##r.c8aa]tRt^>toV3RltRtRtVtV;t#)SwiGLUc$<\SV`4R#N)super__init__)self __class__s&r+rXSwiGLU.__init__?s r.c\W4#rV)rR)rYxgates&&&r+__call__SwiGLU.__call__Bs ar.r/ r0r1r2r3rXr_r5r6 __classcell__rZr*s@@r+rTrT>sr.rTcTaa]tRt^FtoV3RlV3RlltRV3RlRlltRtVtV;t#)AttentionBlockc <V^8dQhRS[/#rconfigr)r)r*s"r+r,AttentionBlock.__annotate__Gs! ! y! r.c<\SV`4VPVnVPVnVPVnVPVP,Vn\ P!VP34Vn\P!VPVPVP,RR7Vn \P!VPVPVP,RR7Vn \P!VPVPVP,RR7Vn\P!VPVP,VPRR7Vn^\ P"!VP4, Vn\'VPVP(RVP*R7VnR#)TbiasF) traditionalscaling_configN)rWrXrrr num_key_value_groupsr;zerossinksnnLinearrq_projk_projv_projo_projmathsqrtsm_scalerr"r#roperYrhrZs&&r+rXAttentionBlock.__init__Gsi  #)#=#= #)#=#=  & &&*D*D D !XXv99;< ii    : :T]] JQU  ii    : :T]] JQU  ii    : :T]] JQU  ii MMF66 68J8JQU  DIIfoo66 # MM   !..   r.ch<V^8dQhRS[PRS[PRS[P/#rr]maskreturnr;array)r)r*s"r+r,rjjs3CC"((C"((C288Cr.c VPwrEpVPpVPpVPV4P WERV4P ^^4p VP V4P WERV4P ^^4p VPV4P WERV4P ^^4p VeMVPWPR7p VPWPR7p VPW4wrM"VPV 4p VPV 4p \WWVPW PR7p VPV P ^^4P WER44#))offset)rrr)shaperr rureshapeswapaxesrvrwr|rupdate_and_fetchr r{rrrx) rYr]rcacheBL_DHkqr?vv_hats &&&& r+r_AttentionBlock.__call__js>''a MM  % % KKN " "1Q / 8 8A > KKN " "1Q / 8 8A > KKN " "1Q / 8 8A >   !LL 1A !LL 1A))!/DAq ! A ! A, !DMMJJ {{5>>!Q/77bABBr.) rrvrrpr rxrur|rrr{rwrVrarcs@@r+rereFs"! ! FCCCr.recPaa]tRt^toV3RlV3RlltV3RlRltRtVtV;t#)MLPBlockc <V^8dQhRS[/#rgri)r)r*s"r+r,MLPBlock.__annotate__s##y#r.cr<\SV`4VPVnVPVnVPVn\ VPVP VP\4RR7Vn\P!VPVPRR7Vn RVn R#)T) input_dims hidden_dims num_experts activationrmrlN) rWrXrrrrrrTexpertsrsrtroutersharding_groupr}s&&r+rXMLPBlock.__init__s !--!'!9!9#)#=#=  ))0000x   ii 2 2F4L4LSWX "r.cN<V^8dQhRS[PRS[P/#)rr]rr)r)r*s"r+r,rs#"((rxxr.cVPe\VP4!V4pVPV4p\W PRR7wr4\ P !VRRR7pVPW4pV\ P!VRR7,pVPRR7pVPe+\ PPW`PR7pV#)N)r?r9T)r9preciser:groupr) rr rrCrr;softmaxr expand_dimssum distributedall_sum)rYr]grindicesexpert_weightsys&& r+r_MLPBlock.__call__s    *d11215A KKN#A)A)AKG"dC LL $ ~B7 7 EErEN    *&&q0C0C&DAr.)rrrrrrrarcs@@r+rrs##"r.rcTaa]tRt^toV3RlV3RlltRV3RlRlltRtVtV;t#)TransformerBlockc <V^8dQhRS[/#rgri)r)r*s"r+r,TransformerBlock.__annotate__s  y r.c$<\SV`4\V4Vn\ V4Vn\ P!VPVP4Vn \ P!VPVP4Vn R#rV) rWrXre self_attnrmlprsRMSNormrrinput_layernormpost_attention_layernormr}s&&r+rXTransformerBlock.__init__sg '/F#!zz&*<*Q>QR(*    3 3) %r.ch<V^8dQhRS[PRS[PRS[P/#rr)r)r*s"r+r,rs.  "(( "(( 288 r.cTpVPV4pVPWV4pWA,pTpVPV4pVPV4pWA,pV#rV)rrrr)rYr]rrresiduals&&&& r+r_TransformerBlock.__call__s\   # NN1E * L  ) )! , HHQK Lr.)rrrrrVrarcs@@r+rrs     r.rcTaa]tRt^toV3RlV3RlltRV3RlRlltRtVtV;t#)GptOssMoeModelc <V^8dQhRS[/#rargsri)r)r*s"r+r,GptOssMoeModel.__annotate__s ? ?Y ?r.c`<\SV`4\P!VPVP 4Vn\P!VP VP4Vn VP;'gRR.VP^,,Vn \VP4Uu.uFp\V4NK upVnVPVnVPP#R4VnVPP#R4VnR#uupi)sliding_attentionfull_attentionN)rWrXrs Embeddingrr embed_tokensrrnormr$rrangerlayersr! window_sizeindexswa_idxga_idx)rYrrrZs&& r+rXGptOssMoeModel.__init__s LL$:J:JKJJt//1B1BC ++**  0  # #q (0*8=T=S=S7TU7T!'-7TU ..''--.AB &&,,-=> Vs<D+c^<V^8dQhRS[PRS[S[P,/#)rinputsinput_embeddings)r;rr)r)r*s"r+r,rs,#288, r.cVeTpMVPV4pVfR.\VP4,p\WBVP,4p\WBVP ,VP R7p\VPW P4Fwrxp V R8XdTMTp V!WJV4pK VPV4pV#)N)rr) rlenrr rrrzipr$r) rYrrrr] full_maskswa_masklayerc layer_typers &&&& r+r_GptOssMoeModel.__call__s  ' A!!&)A =FS--E)!4;;-?@ ( T\\"0@0@ %( U >9HDaq!A%N IIaLr.)rrr$rrrr)NNrarcs@@r+rrs ? ?r.rcaa]tRt^toV3RlV3RlltR V3RlRlltRtR V3RlRllt]R4t ]R 4t R t R t Vt V;t#) Modelc <V^8dQhRS[/#rri)r)r*s"r+r,Model.__annotate__sPPYPr.c<\SV`4WnVPVn\ V4Vn\ P!VPVPRR7Vn R#)FrlN) rWrXrrrmodelrsrtrrlm_head)rYrrZs&&r+rXModel.__init__sH  //#D) yy!1!14??O r.c4<V^8dQhRS[P/#)rrr)r)r*s"r+r,rs77rxx7r.cBVPVPW44#rV)rr)rYrrs&&&r+r_Model.__call__s||DJJv566r.c\;QJd*RVP44F 'gK RM RM!RVP444'dV#/pVP4EFwr4RV9dRV9dRV9dAVP\P 4P R4pVPRR4pRV9dVPRR4p\P!VR,4W#PRR 4&\P!VR,4W#PRR 4&KR V9doRV9dhRV9dAVP\P 4P R4pVPRR4pRV9dVPRR4pWBV&EKIR V9db\P!VR,4W#PR R 4&\P!VR,4W#PR R4&EKRV9dWBVPRR4&EKWBV&EK V#)c3,"TF pRV9xK R#5i)zgate_proj.weightNr/).0r?s& r+ !Model.sanitize..s?1!Q&sTF gate_up_projrm_blocksz.weight_scalesz.scales gate_projup_proj down_projgate_up_proj_biaszgate_proj.biasz up_proj.biasdown_proj_biaszdown_proj.biasr).NNrNNN).rNrr).r).r) anykeysitemsviewr;uint32flattenreplace contiguous)rYweights new_weightsr?rs&& r+sanitizeModel.sanitizes 3? ?333? ? ? ?N MMODA"vQ>ryy)11"5A )Y7A> )Y7AFHmmkNG IInkBCEGMMlOE IIni@A!fAo>ryy)11"5A )Y7A> )Y7A!"A$)MM!H+.II&9;KLMMM!I,/II&9>JK"Q&MNAII&68HIJ!"A=$@r.cX<V^8dQhRS[S[PP,/#)rr)rr;rGroup)r)r*s"r+r,rs#'-'-8BNN$8$89'-r.cT;'g\PP4pVP4pVP 4pVP P EF~p\VPPRVR7VPn \VPPRVR7VPn \VPPRVR7VPn \VPPRVR7VPn VP;PV,un VP;PV,unVPPVPP,VPnVPP VPPV,VPPV^,,VPn\#VP$P&P(RVR7\#VP$P&P*RVR7VP$P&P*;P,V,un\#VP$P&P.RVR7WP$nEK R#)zall-to-sharded)shardingrzsharded-to-allrN)r;rinitsizerankrrrrrurvrwrxrr rprrrrrrrrmrr)rYrNRrs&& r+shard Model.shards..,,. JJL JJLZZ&&E%1&&1A&EOO "&2&&1A&EOO "&2&&1A&EOO "&2&&1A&EOO " OO / /A 5 / OO / /A 5 /33??667 OO 0 %*OO$9$933oo99q5%EOO ! %))++557Gu U %))++557Gu U II   ' ' , , 1 ,  !!))4DE (-II $E'r.c.VPP#rV)rr)rYs&r+r Model.layersCszz   r.c RpV#)c@VPR4'dR^@R^/#R#)r group_sizebitsT)endswith)pathrs&&r+ predicate(Model.quant_predicate..predicateIs#}}X&&$b&!44r.r/)rYrs& r+quant_predicateModel.quant_predicateGs  r.c.pVPPFTpVR8XdVP\44K%VP\ VP P R74KV V#)r)max_size)rr$appendr rrr!)rYcacheslts& r+ make_cacheModel.make_cachePsT**((B%% gi( otyy7O7OPQ )  r.)rrrrrV)r0r1r2r3rXr_rrpropertyrrr%r5r6rbrcs@@r+rrs^PP77%N'-'-R!!r.r)r)gZd;?g@)'ry dataclassesr functoolsrtypingrrmlx.corecorer;mlx.nnrsmlx.nn.layers.distributedrrr baser r r rr r rope_utilsr switch_layersrrrCcompilerRModulerTrerrrrr/r.r+r4s ! PPTT+'$   &' t$ $% $RYY9CRYY9Cx$ryy$Nryy.%RYY%PoBIIor.