From c20106eb62cad09feba8306d3ba1511076c4f37b Mon Sep 17 00:00:00 2001 From: Wing Lian Date: Wed, 25 Jun 2025 08:15:41 -0400 Subject: [PATCH 1/3] fix vocab_size path for gemma3 --- src/liger_kernel/transformers/model/gemma3.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/liger_kernel/transformers/model/gemma3.py b/src/liger_kernel/transformers/model/gemma3.py index 317dc8ebd..42b0ff1b0 100644 --- a/src/liger_kernel/transformers/model/gemma3.py +++ b/src/liger_kernel/transformers/model/gemma3.py @@ -208,7 +208,7 @@ def multimodal_forward( is_training = token_type_ids is not None and labels is not None # Replace image id woth PAD if the image token if OOV, to avoid index-errors - if input_ids is not None and self.config.image_token_index >= self.vocab_size: + if input_ids is not None and self.config.image_token_index >= self.config.text_config.vocab_size: special_image_mask = input_ids == self.config.image_token_index llm_input_ids = input_ids.clone() llm_input_ids[special_image_mask] = 0 From 6e04ee7c40b6109ca0a3ba07a9c451cac6ac52f6 Mon Sep 17 00:00:00 2001 From: Wing Lian Date: Wed, 25 Jun 2025 08:19:36 -0400 Subject: [PATCH 2/3] fix pad_token_id also --- src/liger_kernel/transformers/model/gemma3.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/liger_kernel/transformers/model/gemma3.py b/src/liger_kernel/transformers/model/gemma3.py index 42b0ff1b0..ca30bcd11 100644 --- a/src/liger_kernel/transformers/model/gemma3.py +++ b/src/liger_kernel/transformers/model/gemma3.py @@ -250,7 +250,7 @@ def multimodal_forward( inputs_embeds = inputs_embeds.masked_scatter(special_image_mask, image_features) # mask out pad-token-ids in labels for BC - if labels is not None and self.pad_token_id in labels: + if labels is not None and self.config.pad_token_id in labels: logger.warning_once( "`labels` contains `pad_token_id` which will be masked with `config.ignore_index`. " "You have to mask out `pad_token_id` when preparing `labels`, this behavior will be removed in v.4.46.", From 1bcd24dc2b35819b94cac4b951d87d4c0dbf0169 Mon Sep 17 00:00:00 2001 From: Wing Lian Date: Wed, 25 Jun 2025 08:35:22 -0400 Subject: [PATCH 3/3] more fixes --- .idea/workspace.xml | 45 ++++++++++++------- src/liger_kernel/transformers/model/gemma3.py | 8 ++-- 2 files changed, 34 insertions(+), 19 deletions(-) diff --git a/.idea/workspace.xml b/.idea/workspace.xml index cc065936b..486215dad 100644 --- a/.idea/workspace.xml +++ b/.idea/workspace.xml @@ -4,7 +4,10 @@