From 7c6b296fc130ef81c620cba6f4128bfb0a98e867 Mon Sep 17 00:00:00 2001 From: luoxiaoc Date: Wed, 13 Aug 2025 19:23:52 +0800 Subject: [PATCH 1/4] fix error vocab_size at Qwen2_5_VLForConditionalGeneration loss_function Signed-off-by: luoxiaoc --- src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py b/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py index b4e521ff935e..dbf370b2c7c8 100644 --- a/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py +++ b/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py @@ -749,7 +749,7 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.vocab_size) + loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size) return Qwen2_5_VLCausalLMOutputWithPast( loss=loss, From 8cf5de9799a79850a45bb722a68ae753861bebba Mon Sep 17 00:00:00 2001 From: luoxiaoc Date: Fri, 15 Aug 2025 15:38:26 +0800 Subject: [PATCH 2/4] fix similar errer at qwen2_vl and do make fix-copies Signed-off-by: luoxiaoc --- src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py | 2 +- src/transformers/models/qwen2_vl/modeling_qwen2_vl.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py b/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py index 4d2c00c51ae8..2d8411b59499 100644 --- a/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py +++ b/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py @@ -1524,7 +1524,7 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.vocab_size) + loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size) return Qwen2_5_VLCausalLMOutputWithPast( loss=loss, diff --git a/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py b/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py index 4cae17faece4..7939dd07da13 100644 --- a/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py +++ b/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py @@ -1405,7 +1405,7 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.vocab_size) + loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size) return Qwen2VLCausalLMOutputWithPast( loss=loss, From cbc8208180b6ecc339ab7a8450d846806cef4288 Mon Sep 17 00:00:00 2001 From: luoxiaoc Date: Mon, 18 Aug 2025 16:01:12 +0800 Subject: [PATCH 3/4] pass in kwargs for loss_func at qwen2_vl and qwen2_5_vl Signed-off-by: luoxiaoc --- src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py | 4 +++- src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py | 3 ++- src/transformers/models/qwen2_vl/modeling_qwen2_vl.py | 3 ++- 3 files changed, 7 insertions(+), 3 deletions(-) diff --git a/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py b/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py index 2d8411b59499..90d36320b2fd 100644 --- a/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py +++ b/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py @@ -1524,7 +1524,9 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size) + loss = self.loss_function( + logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, **kwargs + ) return Qwen2_5_VLCausalLMOutputWithPast( loss=loss, diff --git a/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py b/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py index dbf370b2c7c8..8a84d50942bf 100644 --- a/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py +++ b/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py @@ -749,7 +749,8 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size) + loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, + **kwargs) return Qwen2_5_VLCausalLMOutputWithPast( loss=loss, diff --git a/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py b/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py index 7939dd07da13..8711d2aedfe1 100644 --- a/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py +++ b/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py @@ -1405,7 +1405,8 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size) + loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, + **kwargs) return Qwen2VLCausalLMOutputWithPast( loss=loss, From 0b6ce4aba91b8cba0b3373f888d48a6b1c495979 Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" Date: Mon, 18 Aug 2025 08:17:40 +0000 Subject: [PATCH 4/4] Apply style fixes --- src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py | 5 +++-- src/transformers/models/qwen2_vl/modeling_qwen2_vl.py | 5 +++-- 2 files changed, 6 insertions(+), 4 deletions(-) diff --git a/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py b/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py index 8a84d50942bf..1e3f4d87359f 100644 --- a/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py +++ b/src/transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py @@ -749,8 +749,9 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, - **kwargs) + loss = self.loss_function( + logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, **kwargs + ) return Qwen2_5_VLCausalLMOutputWithPast( loss=loss, diff --git a/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py b/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py index 8711d2aedfe1..0df3e9215e21 100644 --- a/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py +++ b/src/transformers/models/qwen2_vl/modeling_qwen2_vl.py @@ -1405,8 +1405,9 @@ def forward( loss = None if labels is not None: - loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, - **kwargs) + loss = self.loss_function( + logits=logits, labels=labels, vocab_size=self.config.text_config.vocab_size, **kwargs + ) return Qwen2VLCausalLMOutputWithPast( loss=loss,