Closed coderkhaleesi closed 4 years ago
Bug description I was trying to run the commands given below for the self-feeding pre-trained model. https://github.com/facebookresearch/ParlAI/tree/master/projects/self_feeding/
Reproduction steps Run python3 projects/self_feeding/interactive.py --model-file zoo:self_feeding/hh131k_hb60k_fb60k_st1k/model --no-cuda --request-feedback true.
Expected behavior The command should run as it is without any errors.
Logs
ubuntu@td-parlai-large:~/ParlAI$ python3 projects/self_feeding/interactive.py --model-file zoo:self_feeding/hh131k_hb60k_fb60k_st1k/model --no-cuda --request-feedback true /home/ubuntu/ParlAI/parlai/agents/transformer/modules.py:32: UserWarning: Installing APEX can give a significant speed boost. warn_once("Installing APEX can give a significant speed boost.") [ warning: overriding opt['subtasks'] to ['dialog', 'satisfaction'] (previously: ['dialog', 'feedback', 'satisfaction'] )] [ warning: overriding opt['interactive'] to True (previously: False )] [ warning: overriding opt['interactive_task'] to True (previously: None )] [ warning: overriding opt['prev_response_filter'] to True (previously: False )] [ warning: overriding opt['partial_load'] to True (previously: False )] [ warning: overriding opt['eval_candidates'] to fixed (previously: inline )] [ warning: overriding opt['encode_candidate_vecs'] to True (previously: False )] [ warning: overriding opt['fixed_candidates_path'] to /home/ubuntu/ParlAI/data/self_feeding/convai2_cands.txt (previously: None )] [ warning: overriding opt['no_cuda'] to True (previously: False )] [ warning: overriding opt['request_feedback'] to True (previously: False )] /home/ubuntu/ParlAI/projects/self_feeding/self_feeding_agent.py:293: UserWarning: Old model: overriding `add_double_person_tokens` to True. warn_once('Old model: overriding `add_double_person_tokens` to True.') [ Setting interactive mode defaults... ] Dictionary: loading dictionary from /home/ubuntu/ParlAI/data/models/self_feeding/hh131k_hb60k_fb60k_st1k/model.dict [ num words = 23617 ] [SelfFeeding: full interactive mode on.] 0%| | 0/1967280 [00:00<?, ?it/s]Skipping token b'1999995' with 1-dimensional vector [b'300']; likely a header 98%|██████████████████████████████████████████████████████████████████████████████████████████████▍ | 1934277/1967280 [08:57<00:07, 4152.70it/s]100%|███████████████████████████████████████████████████████████████████████████████████████████████▉| 1966965/1967280 [09:06<00:00, 3238.55it/s]Traceback (most recent call last): File "projects/self_feeding/interactive.py", line 90, in <module> interactive(parser.parse_args(print_args=False), print_parser=parser) File "projects/self_feeding/interactive.py", line 71, in interactive agent = create_agent(opt, requireModelExists=True) File "/home/ubuntu/ParlAI/parlai/core/agents.py", line 736, in create_agent model = load_agent_module(opt) File "/home/ubuntu/ParlAI/parlai/core/agents.py", line 601, in load_agent_module return model_class(new_opt) File "/home/ubuntu/ParlAI/projects/self_feeding/self_feeding_agent.py", line 256, in __init__ super().__init__(opt, shared) File "/home/ubuntu/ParlAI/parlai/agents/transformer/transformer.py", line 187, in __init__ super().__init__(opt, shared) File "/home/ubuntu/ParlAI/parlai/core/torch_ranker_agent.py", line 170, in __init__ self.model = self.build_model() File "/home/ubuntu/ParlAI/projects/self_feeding/self_feeding_agent.py", line 308, in build_model self._copy_embeddings(embeddings.weight, self.opt['embedding_type']) File "/home/ubuntu/ParlAI/parlai/core/torch_agent.py", line 1198, in _copy_embeddings embs, name = self._get_embtype(emb_type) File "/home/ubuntu/ParlAI/parlai/core/torch_agent.py", line 1136, in _get_embtype embs = download(self.opt.get('datapath')) File "/home/ubuntu/ParlAI/parlai/zoo/fasttext_cc_vectors/build.py", line 21, in download cache=modelzoo_path(datapath, 'models:fasttext_cc_vectors'), File "/home/ubuntu/.local/lib/python3.6/site-packages/torchtext/vocab.py", line 323, in __init__ self.cache(name, cache, url=url, max_vectors=max_vectors) File "/home/ubuntu/.local/lib/python3.6/site-packages/torchtext/vocab.py", line 406, in cache dim)) RuntimeError: Vector for token b'Champia' has 85 dimensions, but previously read vectors have 300 dimensions. All vectors must have the same number of dimensions. 100%|███████████████████████████████████████████████████████████████████████████████████████████████▉| 1966965/1967280 [09:07<00:00, 3595.50it/s]
It looks like your word vectors have failed to download correctly. Run rm -rf data/models/*_vectors and try it again.
rm -rf data/models/*_vectors
Thanks @stephenroller . It worked.
Bug description I was trying to run the commands given below for the self-feeding pre-trained model. https://github.com/facebookresearch/ParlAI/tree/master/projects/self_feeding/
Reproduction steps Run python3 projects/self_feeding/interactive.py --model-file zoo:self_feeding/hh131k_hb60k_fb60k_st1k/model --no-cuda --request-feedback true.
Expected behavior The command should run as it is without any errors.
Logs