bridge.recipes.qwen_vl.qwen25_vl#

Qwen2.5-VL finetuning recipes with parameterless API.

This module provides SFT and PEFT configurations for Qwen2.5-VL models (3B, 7B, 32B, 72B).

Module Contents#

Functions#

qwen25_vl_3b_sft_config

Return a full SFT config for Qwen2.5-VL 3B Instruct.

qwen25_vl_7b_sft_config

Return a full SFT config for Qwen2.5-VL 7B Instruct.

qwen25_vl_32b_sft_config

Return a full SFT config for Qwen2.5-VL 32B Instruct.

qwen25_vl_72b_sft_config

Return a full SFT config for Qwen2.5-VL 72B Instruct.

qwen25_vl_3b_peft_config

Return a PEFT config for Qwen2.5-VL 3B Instruct.

qwen25_vl_7b_peft_config

Return a PEFT config for Qwen2.5-VL 7B Instruct.

qwen25_vl_32b_peft_config

Return a PEFT config for Qwen2.5-VL 32B Instruct.

qwen25_vl_72b_peft_config

Return a PEFT config for Qwen2.5-VL 72B Instruct.

API#

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_3b_sft_config() → megatron.bridge.training.config.ConfigContainer#

Return a full SFT config for Qwen2.5-VL 3B Instruct.

Default configuration: 1 node, 8 GPUs

  • TP=1, PP=1

  • LR=5e-6 (full SFT)

  • Sequence length: 4096

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_7b_sft_config() → megatron.bridge.training.config.ConfigContainer#

Return a full SFT config for Qwen2.5-VL 7B Instruct.

Default configuration: 1 node, 8 GPUs

  • TP=2, PP=1

  • LR=5e-6 (full SFT)

  • Sequence length: 4096

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_32b_sft_config() → megatron.bridge.training.config.ConfigContainer#

Return a full SFT config for Qwen2.5-VL 32B Instruct.

Default configuration: 2 nodes, 16 GPUs total

  • TP=8, PP=2

  • LR=5e-6 (full SFT)

  • Sequence length: 4096

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_72b_sft_config() → megatron.bridge.training.config.ConfigContainer#

Return a full SFT config for Qwen2.5-VL 72B Instruct.

Default configuration: 4 nodes, 32 GPUs total

  • TP=8, PP=4

  • LR=5e-6 (full SFT)

  • Sequence length: 4096

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_3b_peft_config(
peft_scheme: str | megatron.bridge.peft.base.PEFT = 'lora',
) → megatron.bridge.training.config.ConfigContainer#

Return a PEFT config for Qwen2.5-VL 3B Instruct.

Default configuration: 1 node, 8 GPUs

  • TP=1, PP=1

  • LR=1e-4 (PEFT)

  • Sequence length: 4096

Parameters:

peft_scheme – PEFT scheme - “lora”, “dora”, or a custom PEFT instance.

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_7b_peft_config(
peft_scheme: str | megatron.bridge.peft.base.PEFT = 'lora',
) → megatron.bridge.training.config.ConfigContainer#

Return a PEFT config for Qwen2.5-VL 7B Instruct.

Default configuration: 1 node, 8 GPUs

  • TP=1, PP=1

  • LR=1e-4 (PEFT)

  • Sequence length: 4096

Parameters:

peft_scheme – PEFT scheme - “lora”, “dora”, or a custom PEFT instance.

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_32b_peft_config(
peft_scheme: str | megatron.bridge.peft.base.PEFT = 'lora',
) → megatron.bridge.training.config.ConfigContainer#

Return a PEFT config for Qwen2.5-VL 32B Instruct.

Default configuration: 1 node, 8 GPUs

  • TP=1, PP=1

  • LR=1e-4 (PEFT)

  • Sequence length: 4096

Parameters:

peft_scheme – PEFT scheme - “lora”, “dora”, or a custom PEFT instance.

bridge.recipes.qwen_vl.qwen25_vl.qwen25_vl_72b_peft_config(
peft_scheme: str | megatron.bridge.peft.base.PEFT = 'lora',
) → megatron.bridge.training.config.ConfigContainer#

Return a PEFT config for Qwen2.5-VL 72B Instruct.

Default configuration: 1 node, 8 GPUs

  • TP=1, PP=1

  • LR=1e-4 (PEFT)

  • Sequence length: 4096

Parameters:

peft_scheme – PEFT scheme - “lora”, “dora”, or a custom PEFT instance.