You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I finetune InternLM-Xcomposer-2.0 with the official finetune script. However, I get the below warning when loading the finetuned model:
Some weights of InternLMXComposer2ForCausalLM were not initialized from the model checkpoint at /nas/data/jkliew/IXL-2-VL-finetuned and are newly initialized: ['model.layers.0.attention.wo.Plora_A.weight', 'model.layers.0.attention.wo.Plora_B.weight', 'model.layers.0.attention.wo.weight', 'model.layers.0.attention.wqkv.Plora_A.weight', 'model.layers.0.attention.wqkv.Plora_B.weight', 'model.layers.0.attention.wqkv.weight', 'model.layers.0.attention_norm.weight', 'model.layers.0.feed_forward.w1.Plora_A.weight', 'model.layers.0.feed_forward.w1.Plora_B.weight', 'model.layers.0.feed_forward.w1.weight', 'model.layers.0.feed_forward.w2.Plora_A.weight', 'model.layers.0.feed_forward.w2.Plora_B.weight' .....]
The output of this fine-tuned model is all random chars with no actual meaning, what's causing this error?
The text was updated successfully, but these errors were encountered:
I finetune InternLM-Xcomposer-2.0 with the official finetune script. However, I get the below warning when loading the finetuned model:
Some weights of InternLMXComposer2ForCausalLM were not initialized from the model checkpoint at /nas/data/jkliew/IXL-2-VL-finetuned and are newly initialized: ['model.layers.0.attention.wo.Plora_A.weight', 'model.layers.0.attention.wo.Plora_B.weight', 'model.layers.0.attention.wo.weight', 'model.layers.0.attention.wqkv.Plora_A.weight', 'model.layers.0.attention.wqkv.Plora_B.weight', 'model.layers.0.attention.wqkv.weight', 'model.layers.0.attention_norm.weight', 'model.layers.0.feed_forward.w1.Plora_A.weight', 'model.layers.0.feed_forward.w1.Plora_B.weight', 'model.layers.0.feed_forward.w1.weight', 'model.layers.0.feed_forward.w2.Plora_A.weight', 'model.layers.0.feed_forward.w2.Plora_B.weight' .....]
The output of this fine-tuned model is all random chars with no actual meaning, what's causing this error?
The text was updated successfully, but these errors were encountered: