batra-mlp-lab / visdial-rl Goto Github PK

This is related to #2
I ran
python -m visdom.server -p 8097
in order to use visdom.

After that, I ran
python evaluate.py -useGPU \ -startFrom checkpoints/abot_sl_ep60.vd \ -qstartFrom checkpoints/qbot_sl_ep60.vd \ -evalMode ABotRank QBotRank \ -visdomServer http://127.0.0.1 \ -visdomServerPort 8097 \ -visdomEnv ABotRank-QBotRank
Even though I clearly designated which server to use and the server port number, somehow they are overwritten as follows. How can I fix this?

{'CELossCoeff': 200,
'batchSize': 20,
'beamSize': 1,
'ckpt_iterid': 152160,
'ckpt_lRate': 4.999614155150613e-05,
'cocoDir': '',
'cocoInfo': '',
'continue': True,
'decoder': 'gen',
'dropout': 0.0,
'embedSize': 300,
'enableVisdom': 1,
'encoder': 'hre-ques-lateim-hist',
'evalModeList': ['ABotRank', 'QBotRank'],
'evalSplit': 'val',
'evalTitle': 'eval',
'featLossCoeff': 1000,
'freezeQFeatNet': 0,
'imgEmbedSize': 300,
'imgFeatureSize': 4096,
'imgNorm': 1,
'inputImg': 'data/visdial/data_img.h5',
'inputJson': 'data/visdial/chat_processed_params.json',
'inputQues': 'data/visdial/chat_processed_data.h5',
'learningRate': 0.001,
'lrDecayRate': 0.9997592083,
'minLRate': 5e-05,
'numEpochs': 65,
'numLayers': 2,
'numRounds': 10,
'numWorkers': 2,
'qdecoder': 'gen',
'qencoder': 'hre-ques-lateim-hist',
'qstartFrom': 'checkpoints/qbot_sl_ep60.vd',
'randomSeed': 32,
'rlAbotReward': 1,
'rnnHiddenSize': 512,
'saveName': 'debug-final5_sl-qbot',
'savePath': 'checkpoints/debug-final5_sl-qbot',
'startFrom': 'checkpoints/abot_sl_ep60.vd',
'trainMode': 'sl-abot',
'useCurriculum': 1,
'useGPU': True,
'useHistory': True,
'useIm': 'late',
'verbose': 1,
'visdomEnv': 'ABotRank-QBotRank',
'visdomServer': 'http://walle.cc.gatech.edu',
'visdomServerPort': 8893,
'vocabSize': 7826}

Thanks to your kindness, I managed to run your code.
By the way, here is one more question.
I ran

python evaluate.py -useGPU \
    -startFrom checkpoints/abot_rl_ep20.vd \
    -qstartFrom checkpoints/qbot_rl_ep20.vd \
    -evalMode dialog \
    -cocoDir /my/path/to/coco/images/ \
    -cocoInfo /my/path/to/coco.json \
    -beamSize 5

then implemented

cd dialog_output/
python -m http.server 8000

however, I found that the visualized captions were quite different from those on your pic
There were so many "UNK" in my result. Is it natural? Or not?
And can you tell me in what condition I could make similar results to yours?

Following the paper, the above should be replaced by
loss += -1 * log_prob * (reward.detach() * (self.mask[:, t].float()))

Not having a .detach() on the reward here provides another source of gradients to the feature regression module in addition to the feature loss, the only difference being these gradients are scaled by the log-probs, which does not seem to mean anything intuitively.

Hi,

Thank you providing this code.

Could you explain your code for backtracking in beam search.

In particular how do you handle the dropped sequences that have seen EOS earlier during forward phase as done in this implementation.

I ran the command "python train.py -useGPU -trainMode sl-abot". However, I got such error "RuntimeError: CUDA error: out of memory". How much GPU memory is needed when training the model? The memory of my GPU is 12212MiB. Thanks!

When I tried to run evaluate.py with “-evalMode dialog” like this
python evaluate.py -useGPU \ -startFrom checkpoints/abot_rl_ep20.vd \ -qstartFrom checkpoints/qbot_rl_ep20.vd \ -evalMode dialog \ -beamSize 5
, I had an error as follows

[Error] Need coco directory and info as input to -cocoDir and -cocoInfo arguments for locating coco image files.
Exiting dialogDump without saving files.

What should I do with this? Is it appropriate for me to make a path to the directory which has MSCOCO dataset?

In addition what are these two(-cocoDir, -cocoInfo) used for?

Hello.
Thank you for sharing your great work!

After running
python evaluate.py -useGPU \ -startFrom checkpoints/abot_sl_ep60.vd \ -qstartFrom checkpoints/qbot_sl_ep60.vd \ -evalMode ABotRank QBotRank>
, I had lots of the following error

[Errno110] Connection timed out

I guess before running
python evaluate.py -useGPU \ -startFrom checkpoints/abot_sl_ep60.vd \ -qstartFrom checkpoints/qbot_sl_ep60.vd \ -evalMode ABotRank QBotRank
, I should run
python -m visdom.server -p <port>
in order to visualize the evaluation result.
Is it correct?

If not, how should I deal with this code?

kindly help me error is occur during executing of prepro.py file error is
Traceback (most recent call last):
File "prepro.py", line 274, in
captions_train, captions_train_len, questions_train, questions_train_len, answers_train, answers_train_len, options_train, options_train_list, options_train_len, answers_train_index, images_train_index, images_train_list, _ = create_data_mats(data_train_toks, ques_train_inds, ans_train_inds, args, 'train')
File "prepro.py", line 127, in create_data_mats
options[i][j] = np.array(data_toks[image_id]['dialog'][j]['answer_options']) + 1
ValueError: could not broadcast input array from shape (104) into shape (100)

Do we really need transpose？Because I got this error:
beamTokensTable[:, :, 0] = topIdx.transpose(0, 1).data
RuntimeError: The expanded size of the tensor (5) must match the existing size (10) at non-singleton dimension 1. Target sizes: [10, 5]. Tensor sizes: [5, 10]

Hi
When I run this command

CUDA_VISIBLE_DEVICES=4 python train.py -useGPU -trainMode sl-qbot -enableVisdom 1
-visdomServer http://127.0.0.1 -visdomServerPort 8098 -visdomEnv my-qbot-job

I got this error

File "train.py", line 39, in
dataset = VisDialDataset(params, splits)
File "/home/badri/badripatro/workspace_project/visual_dialog/visdial-rl-master_v10/dataloader.py", line 149, in init
self.prepareDataset(dtype)
File "/visdial-rl-master_v10/dataloader.py", line 185, in prepareDataset
self.data[dtype + '_ans_ind'] -= 1
KeyError: 'test_ans_ind'

Should I change in the "propro.py" file?

These are the keys present in "chat_processed_data.h5" file

f = h5py.File("chat_processed_data.h5")
list(f)
['ans_index_train', 'ans_index_val', 'ans_length_test', 'ans_length_train', 'ans_length_val', 'ans_test', 'ans_train', 'ans_val', 'cap_length_test', 'cap_length_train', 'cap_length_val', 'cap_test', 'cap_train', 'cap_val', 'img_pos_test', 'img_pos_train', 'img_pos_val', 'num_rounds_test', 'opt_length_test', 'opt_length_train', 'opt_length_val', 'opt_list_test', 'opt_list_train', 'opt_list_val', 'opt_test', 'opt_train', 'opt_val', 'ques_length_test', 'ques_length_train', 'ques_length_val', 'ques_test', 'ques_train', 'ques_val']