RWKV is a RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.
Apache License 2.0
406
stars
40
forks
source link
微调后运行,AttributeError: 'types.SimpleNamespace' object has no attribute 'emb' #14
Run prompt...
Traceback (most recent call last):
File "/home/v/下载/RWKV-LM-LoRA-main/RWKV-v4neo/chat.py", line 187, in
out = run_rnn(tokenizer.tokenizer.encode(init_prompt))
File "/home/v/下载/RWKV-LM-LoRA-main/RWKV-v4neo/chat.py", line 157, in run_rnn
current_state = model.forward(model_tokens, current_state, preprocess_only = True)
File "/home/v/下载/RWKV-LM-LoRA-main/RWKV-v4neo/src/model_run.py", line 225, in forward
x = w.emb.weight[ctx[-1]]
AttributeError: 'types.SimpleNamespace' object has no attribute 'emb'
v@v-virtual-machine:~/下载/RWKV-LM-LoRA-main/RWKV-v4neo$
Run prompt... Traceback (most recent call last): File "/home/v/下载/RWKV-LM-LoRA-main/RWKV-v4neo/chat.py", line 187, in
out = run_rnn(tokenizer.tokenizer.encode(init_prompt))
File "/home/v/下载/RWKV-LM-LoRA-main/RWKV-v4neo/chat.py", line 157, in run_rnn
current_state = model.forward(model_tokens, current_state, preprocess_only = True)
File "/home/v/下载/RWKV-LM-LoRA-main/RWKV-v4neo/src/model_run.py", line 225, in forward
x = w.emb.weight[ctx[-1]]
AttributeError: 'types.SimpleNamespace' object has no attribute 'emb'
v@v-virtual-machine:~/下载/RWKV-LM-LoRA-main/RWKV-v4neo$