From 8942b2fc76c59069af9b79acff7b1336bccad1f5 Mon Sep 17 00:00:00 2001 From: Daniel Han Date: Thu, 31 Oct 2024 01:23:41 -0700 Subject: [PATCH] Update cross_entropy_loss.py --- unsloth/kernels/cross_entropy_loss.py | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/unsloth/kernels/cross_entropy_loss.py b/unsloth/kernels/cross_entropy_loss.py index 8cbdbf2a2c..a13475e227 100644 --- a/unsloth/kernels/cross_entropy_loss.py +++ b/unsloth/kernels/cross_entropy_loss.py @@ -92,10 +92,10 @@ def _cross_entropy_forward( pass -# @triton.heuristics({ -# "DO_SOFTCAPPING": lambda args: args["DO_SOFTCAPPING" ], -# "DO_LOGIT_SCALING": lambda args: args["DO_LOGIT_SCALING"], -# }) +@triton.heuristics({ + "DO_SOFTCAPPING": lambda args: args["DO_SOFTCAPPING" ], + "DO_LOGIT_SCALING": lambda args: args["DO_LOGIT_SCALING"], +}) @triton.jit def _chunked_cross_entropy_forward( logits_ptr , @@ -106,9 +106,9 @@ def _chunked_cross_entropy_forward( VOCAB_SIZE , N_CHUNKS , BLOCK_SIZE : tl.constexpr, - DO_SOFTCAPPING , + DO_SOFTCAPPING : tl.constexpr(tl.int1), SOFTCAP , - DO_LOGIT_SCALING , + DO_LOGIT_SCALING : tl.constexpr(tl.int1), LOGIT_SCALE , ): """ @@ -189,9 +189,9 @@ def _cross_entropy_backward( labels_ptr , VOCAB_SIZE , BLOCK_SIZE : tl.constexpr, - DO_SOFTCAPPING , + DO_SOFTCAPPING : tl.constexpr(tl.int1), SOFTCAP , - DO_LOGIT_SCALING , + DO_LOGIT_SCALING : tl.constexpr(tl.int1), LOGIT_SCALE , ): """ @@ -347,7 +347,7 @@ class Fast_CrossEntropyLoss(torch.autograd.Function): SOFTCAP = ctx.logit_softcapping, DO_LOGIT_SCALING = ctx.DO_LOGIT_SCALING, LOGIT_SCALE = ctx.logit_scaling, - num_warps = 8, + num_warps = 8, ) return logits, None, None, None, pass