Open AmuzeLu opened 3 hours ago
[2024-09-21 09:32:21,101] [INFO] [real_accelerator.py:203:get_accelerator] Setting ds_accelerator to cuda (auto detect)
llamafactory
llamafactory-cli export examples/merge_lora/qwen2_vl_lora_sft.yaml
Lora微调没有问题,但是在执行合并的过程中出现了如下问题。
ValueError: Unrecognized configuration class <class 'transformers.models.qwen2_vl.configuration_qwen2_vl.Qwen2VLConfig'> for this kind of AutoModel: AutoModelForCausalLM. Model type should be one of BartConfig, BertConfig, BertGenerationConfig, BigBirdConfig, BigBirdPegasusConfig, BioGptConfig, BlenderbotConfig, BlenderbotSmallConfig, BloomConfig, CamembertConfig, LlamaConfig, CodeGenConfig, CohereConfig, CpmAntConfig, CTRLConfig, Data2VecTextConfig, DbrxConfig, ElectraConfig, ErnieConfig, FalconConfig, FalconMambaConfig, FuyuConfig, GemmaConfig, Gemma2Config, GitConfig, GPT2Config, GPT2Config, GPTBigCodeConfig, GPTNeoConfig, GPTNeoXConfig, GPTNeoXJapaneseConfig, GPTJConfig, GraniteConfig, JambaConfig, JetMoeConfig, LlamaConfig, MambaConfig, Mamba2Config, MarianConfig, MBartConfig, MegaConfig, MegatronBertConfig, MistralConfig, MixtralConfig, MptConfig, MusicgenConfig, MusicgenMelodyConfig, MvpConfig, NemotronConfig, OlmoConfig, OlmoeConfig, OpenLlamaConfig, OpenAIGPTConfig, OPTConfig, PegasusConfig, PersimmonConfig, PhiConfig, Phi3Config, PLBartConfig, ProphetNetConfig, QDQBertConfig, Qwen2Config, Qwen2MoeConfig, RecurrentGemmaConfig, ReformerConfig, RemBertConfig, RobertaConfig, RobertaPreLayerNormConfig, RoCBertConfig, RoFormerConfig, RwkvConfig, Speech2Text2Config, StableLmConfig, Starcoder2Config, TransfoXLConfig, TrOCRConfig, WhisperConfig, XGLMConfig, XLMConfig, XLMProphetNetConfig, XLMRobertaConfig, XLMRobertaXLConfig, XLNetConfig, XmodConfig.
No response
add visual_inputs: true in yaml files will solve the problem as shown in #5316.
visual_inputs: true
Recommend that add visual_inputs: true to examples/merge_lora/qwen2vl_lora_sft.yaml.
Reminder
System Info
[2024-09-21 09:32:21,101] [INFO] [real_accelerator.py:203:get_accelerator] Setting ds_accelerator to cuda (auto detect)
llamafactory
version: 0.8.4.dev0Reproduction
llamafactory-cli export examples/merge_lora/qwen2_vl_lora_sft.yaml
Lora微调没有问题,但是在执行合并的过程中出现了如下问题。
Expected behavior
No response
Others
No response