search

isayev / ReLeaSE

Deep Reinforcement Learning for de-novo Drug Design
MIT License
352 stars 135 forks source link

LogP Example: "TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list" #38

Open gmseabra opened 4 years ago

gmseabra commented 4 years ago

Hi,

I'm re-running the LogP example using current version of PyTorch, and the execution stops in the reinforcement loop due to a TypeError, as below. Are you aware of any changes in PyTorch that could be responsible for this? Is there a solution for it?

Thanks!

for i in range(n_iterations):
    for j in trange(n_policy, desc='Policy gradient...'):
        cur_reward, cur_loss = RL_logp.policy_gradient(gen_data)
        rewards.append(simple_moving_average(rewards, cur_reward)) 
        rl_losses.append(simple_moving_average(rl_losses, cur_loss))

    plt.plot(rewards)
    plt.xlabel('Training iteration')
    plt.ylabel('Average reward')
    plt.show()
    plt.plot(rl_losses)
    plt.xlabel('Training iteration')
    plt.ylabel('Loss')
    plt.show()

    smiles_cur, prediction_cur = estimate_and_update(RL_logp.generator, 
                                                     my_predictor, 
                                                     n_to_generate)
    print('Sample trajectories:')
    for sm in smiles_cur[:5]:
        print(sm)

with the error below:

Policy gradient...:   0%|          | 0/15 [00:00<?, ?it/s]./release/data.py:98: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  return torch.tensor(tensor).cuda()
Policy gradient...:   0%|          | 0/15 [00:00<?, ?it/s]

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-37-7a3a9698cf0c> in <module>
      1 for i in range(n_iterations):
      2     for j in trange(n_policy, desc='Policy gradient...'):
----> 3         cur_reward, cur_loss = RL_logp.policy_gradient(gen_data)
      4         rewards.append(simple_moving_average(rewards, cur_reward))
      5         rl_losses.append(simple_moving_average(rl_losses, cur_loss))

~/work/li/leadopt/generator/ReLeaSE/release/reinforcement.py in policy_gradient(self, data, n_batch, gamma, std_smiles, grad_clipping, **kwargs)
    117                     reward = self.get_reward(trajectory[1:-1],
    118                                              self.predictor,
--> 119                                              **kwargs)
    120 
    121             # Converting string of characters into tensor

<ipython-input-33-a8c049e9e937> in get_reward_logp(smiles, predictor, invalid_reward)
      1 def get_reward_logp(smiles, predictor, invalid_reward=0.0):
----> 2     mol, prop, nan_smiles = predictor.predict([smiles])
      3     if len(nan_smiles) == 1:
      4         return invalid_reward
      5     if (prop[0] >= 1.0) and (prop[0] <= 4.0):

~/work/li/leadopt/generator/ReLeaSE/release/rnn_predictor.py in predict(self, smiles, use_tqdm)
     62                 self.model[i]([torch.LongTensor(smiles_tensor).cuda(),
     63                                torch.LongTensor(length).cuda()],
---> 64                               eval=True).detach().cpu().numpy())
     65         prediction = np.array(prediction).reshape(len(self.model), -1)
     66         prediction = np.min(prediction, axis=0)

/opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    720             result = self._slow_forward(*input, **kwargs)
    721         else:
--> 722             result = self.forward(*input, **kwargs)
    723         for hook in itertools.chain(
    724                 _global_forward_hooks.values(),

~/work/source/repos/OpenChem/openchem/models/Smiles2Label.py in forward(self, inp, eval)
     41         else:
     42             self.train()
---> 43         embedded = self.Embedding(inp)
     44         output, _ = self.Encoder(embedded)
     45         output = self.MLP(output)

/opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    720             result = self._slow_forward(*input, **kwargs)
    721         else:
--> 722             result = self.forward(*input, **kwargs)
    723         for hook in itertools.chain(
    724                 _global_forward_hooks.values(),

~/work/source/repos/OpenChem/openchem/modules/embeddings/basic_embedding.py in forward(self, inp)
      7 
      8     def forward(self, inp):
----> 9         embedded = self.embedding(inp)
     10         return embedded

/opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    720             result = self._slow_forward(*input, **kwargs)
    721         else:
--> 722             result = self.forward(*input, **kwargs)
    723         for hook in itertools.chain(
    724                 _global_forward_hooks.values(),

/opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/sparse.py in forward(self, input)
    124         return F.embedding(
    125             input, self.weight, self.padding_idx, self.max_norm,
--> 126             self.norm_type, self.scale_grad_by_freq, self.sparse)
    127 
    128     def extra_repr(self) -> str:

/opt/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py in embedding(input, weight, padding_idx, max_norm, norm_type, scale_grad_by_freq, sparse)
   1812         # remove once script supports set_grad_enabled
   1813         _no_grad_embedding_renorm_(weight, input, max_norm, norm_type)
-> 1814     return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)
   1815 
   1816 

TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list
isayev commented 4 years ago

Gustavo, I guess your data is a list, but the model expects a tensor.

On Mon, Oct 19, 2020 at 4:48 PM Gustavo Seabra notifications@github.com wrote:

Hi,

I'm re-running the LogP example using current version of PyTorch, and the execution stops in the reinforcement loop due to a TypeError, as below. Are you aware of any changes in PyTorch that could be responsible for this? Is there a solution for it?

Thanks!

for i in range(n_iterations): for j in trange(n_policy, desc='Policy gradient...'): cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) rewards.append(simple_moving_average(rewards, cur_reward)) rl_losses.append(simple_moving_average(rl_losses, cur_loss))

plt.plot(rewards)
plt.xlabel('Training iteration')
plt.ylabel('Average reward')
plt.show()
plt.plot(rl_losses)
plt.xlabel('Training iteration')
plt.ylabel('Loss')
plt.show()

smiles_cur, prediction_cur = estimate_and_update(RL_logp.generator,
                                                 my_predictor,
                                                 n_to_generate)
print('Sample trajectories:')
for sm in smiles_cur[:5]:
    print(sm)

with the error below:

Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]./release/data.py:98: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requiresgrad(True), rather than torch.tensor(sourceTensor). return torch.tensor(tensor).cuda() Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]

TypeError Traceback (most recent call last)

in 1 for i in range(n_iterations): 2 for j in trange(n_policy, desc='Policy gradient...'): ----> 3 cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) 4 rewards.append(simple_moving_average(rewards, cur_reward)) 5 rl_losses.append(simple_moving_average(rl_losses, cur_loss)) ~/work/li/leadopt/generator/ReLeaSE/release/reinforcement.py in policy_gradient(self, data, n_batch, gamma, std_smiles, grad_clipping, **kwargs) 117 reward = self.get_reward(trajectory[1:-1], 118 self.predictor, --> 119 **kwargs) 120 121 # Converting string of characters into tensor in get_reward_logp(smiles, predictor, invalid_reward) 1 def get_reward_logp(smiles, predictor, invalid_reward=0.0): ----> 2 mol, prop, nan_smiles = predictor.predict([smiles]) 3 if len(nan_smiles) == 1: 4 return invalid_reward 5 if (prop[0] >= 1.0) and (prop[0] <= 4.0): ~/work/li/leadopt/generator/ReLeaSE/release/rnn_predictor.py in predict(self, smiles, use_tqdm) 62 self.model[i]([torch.LongTensor(smiles_tensor).cuda(), 63 torch.LongTensor(length).cuda()], ---> 64 eval=True).detach().cpu().numpy()) 65 prediction = np.array(prediction).reshape(len(self.model), -1) 66 prediction = np.min(prediction, axis=0) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/models/Smiles2Label.py in forward(self, inp, eval) 41 else: 42 self.train() ---> 43 embedded = self.Embedding(inp) 44 output, _ = self.Encoder(embedded) 45 output = self.MLP(output) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/modules/embeddings/basic_embedding.py in forward(self, inp) 7 8 def forward(self, inp): ----> 9 embedded = self.embedding(inp) 10 return embedded /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/sparse.py in forward(self, input) 124 return F.embedding( 125 input, self.weight, self.padding_idx, self.max_norm, --> 126 self.norm_type, self.scale_grad_by_freq, self.sparse) 127 128 def extra_repr(self) -> str: /opt/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py in embedding(input, weight, padding_idx, max_norm, norm_type, scale_grad_by_freq, sparse) 1812 # remove once script supports set_grad_enabled 1813 _no_grad_embedding_renorm_(weight, input, max_norm, norm_type) -> 1814 return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse) 1815 1816 TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub , or unsubscribe .
gmseabra commented 4 years ago

Well... I'm no wiz, but that much I had figured out.

The point is that there's no such thing as "my data": I'm just running the LogP Jupyter notebook from the git repo.

I assume it did work fine when it was created, with PyTorch 0.4. But maybe there was some change in PyTorch internals?

-- Gustavo Seabra

From: Olexandr Isayev notifications@github.com Sent: Monday, October 19, 2020 6:51:58 PM To: isayev/ReLeaSE ReLeaSE@noreply.github.com Cc: de Miranda Seabra, Gustavo seabra@cop.ufl.edu; Author author@noreply.github.com Subject: Re: [isayev/ReLeaSE] LogP Example: "TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list" (#38)

[External Email]

Gustavo, I guess your data is a list, but the model expects a tensor.

On Mon, Oct 19, 2020 at 4:48 PM Gustavo Seabra notifications@github.com wrote:

Hi,

I'm re-running the LogP example using current version of PyTorch, and the execution stops in the reinforcement loop due to a TypeError, as below. Are you aware of any changes in PyTorch that could be responsible for this? Is there a solution for it?

Thanks!

for i in range(n_iterations): for j in trange(n_policy, desc='Policy gradient...'): cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) rewards.append(simple_moving_average(rewards, cur_reward)) rl_losses.append(simple_moving_average(rl_losses, cur_loss))

plt.plot(rewards) plt.xlabel('Training iteration') plt.ylabel('Average reward') plt.show() plt.plot(rl_losses) plt.xlabel('Training iteration') plt.ylabel('Loss') plt.show()

smiles_cur, prediction_cur = estimate_and_update(RL_logp.generator, my_predictor, n_to_generate) print('Sample trajectories:') for sm in smiles_cur[:5]: print(sm)

with the error below:

Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]./release/data.py:98: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requiresgrad(True), rather than torch.tensor(sourceTensor). return torch.tensor(tensor).cuda() Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]

TypeError Traceback (most recent call last)

in 1 for i in range(n_iterations): 2 for j in trange(n_policy, desc='Policy gradient...'): ----> 3 cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) 4 rewards.append(simple_moving_average(rewards, cur_reward)) 5 rl_losses.append(simple_moving_average(rl_losses, cur_loss)) ~/work/li/leadopt/generator/ReLeaSE/release/reinforcement.py in policy_gradient(self, data, n_batch, gamma, std_smiles, grad_clipping, **kwargs) 117 reward = self.get_reward(trajectory[1:-1], 118 self.predictor, --> 119 **kwargs) 120 121 # Converting string of characters into tensor in get_reward_logp(smiles, predictor, invalid_reward) 1 def get_reward_logp(smiles, predictor, invalid_reward=0.0): ----> 2 mol, prop, nan_smiles = predictor.predict([smiles]) 3 if len(nan_smiles) == 1: 4 return invalid_reward 5 if (prop[0] >= 1.0) and (prop[0] <= 4.0): ~/work/li/leadopt/generator/ReLeaSE/release/rnn_predictor.py in predict(self, smiles, use_tqdm) 62 self.model[i]([torch.LongTensor(smiles_tensor).cuda(), 63 torch.LongTensor(length).cuda()], ---> 64 eval=True).detach().cpu().numpy()) 65 prediction = np.array(prediction).reshape(len(self.model), -1) 66 prediction = np.min(prediction, axis=0) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/models/Smiles2Label.py in forward(self, inp, eval) 41 else: 42 self.train() ---> 43 embedded = self.Embedding(inp) 44 output, _ = self.Encoder(embedded) 45 output = self.MLP(output) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/modules/embeddings/basic_embedding.py in forward(self, inp) 7 8 def forward(self, inp): ----> 9 embedded = self.embedding(inp) 10 return embedded /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/sparse.py in forward(self, input) 124 return F.embedding( 125 input, self.weight, self.padding_idx, self.max_norm, --> 126 self.norm_type, self.scale_grad_by_freq, self.sparse) 127 128 def extra_repr(self) -> str: /opt/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py in embedding(input, weight, padding_idx, max_norm, norm_type, scale_grad_by_freq, sparse) 1812 # remove once script supports set_grad_enabled 1813 _no_grad_embedding_renorm_(weight, input, max_norm, norm_type) -> 1814 return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse) 1815 1816 TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub , or unsubscribe .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-23issuecomment-2D712484495&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=gRSQ-wfSd5aALN-7K1HnwMMTzxf0A6KsZEqKnW0OxEw&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AKBAWSFN5DXOXZWQJ4PDCADSLS7I5ANCNFSM4SWVLIVA&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=1Jpulkaf66vxDvRV9S4I6Dh3GvJgMBjbwZ5v5UzGI6U&e=.

isayev commented 4 years ago

Oh makes sense. Yeah, it does not work with the latest pytorch, you still have to run it with the old one.

On Mon, Oct 19, 2020 at 8:10 PM Gustavo Seabra notifications@github.com wrote:

Well... I'm no wiz, but that much I had figured out.

The point is that there's no such thing as "my data": I'm just running the LogP Jupyter notebook from the git repo.

I assume it did work fine when it was created, with PyTorch 0.4. But maybe there was some change in PyTorch internals?

-- Gustavo Seabra

From: Olexandr Isayev notifications@github.com Sent: Monday, October 19, 2020 6:51:58 PM To: isayev/ReLeaSE ReLeaSE@noreply.github.com Cc: de Miranda Seabra, Gustavo seabra@cop.ufl.edu; Author < author@noreply.github.com> Subject: Re: [isayev/ReLeaSE] LogP Example: "TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list" (#38)

[External Email]

Gustavo, I guess your data is a list, but the model expects a tensor.

On Mon, Oct 19, 2020 at 4:48 PM Gustavo Seabra notifications@github.com wrote:

Hi,

I'm re-running the LogP example using current version of PyTorch, and the execution stops in the reinforcement loop due to a TypeError, as below. Are you aware of any changes in PyTorch that could be responsible for this? Is there a solution for it?

Thanks!

for i in range(n_iterations): for j in trange(n_policy, desc='Policy gradient...'): cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) rewards.append(simple_moving_average(rewards, cur_reward)) rl_losses.append(simple_moving_average(rl_losses, cur_loss))

plt.plot(rewards) plt.xlabel('Training iteration') plt.ylabel('Average reward') plt.show() plt.plot(rl_losses) plt.xlabel('Training iteration') plt.ylabel('Loss') plt.show()

smiles_cur, prediction_cur = estimate_and_update(RL_logp.generator, my_predictor, n_to_generate) print('Sample trajectories:') for sm in smiles_cur[:5]: print(sm)

with the error below:

Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]./release/data.py:98: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requiresgrad(True), rather than torch.tensor(sourceTensor). return torch.tensor(tensor).cuda() Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]

TypeError Traceback (most recent call last)

in 1 for i in range(n_iterations): 2 for j in trange(n_policy, desc='Policy gradient...'): ----> 3 cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) 4 rewards.append(simple_moving_average(rewards, cur_reward)) 5 rl_losses.append(simple_moving_average(rl_losses, cur_loss)) ~/work/li/leadopt/generator/ReLeaSE/release/reinforcement.py in policy_gradient(self, data, n_batch, gamma, std_smiles, grad_clipping, **kwargs) 117 reward = self.get_reward(trajectory[1:-1], 118 self.predictor, --> 119 **kwargs) 120 121 # Converting string of characters into tensor in get_reward_logp(smiles, predictor, invalid_reward) 1 def get_reward_logp(smiles, predictor, invalid_reward=0.0): ----> 2 mol, prop, nan_smiles = predictor.predict([smiles]) 3 if len(nan_smiles) == 1: 4 return invalid_reward 5 if (prop[0] >= 1.0) and (prop[0] <= 4.0): ~/work/li/leadopt/generator/ReLeaSE/release/rnn_predictor.py in predict(self, smiles, use_tqdm) 62 self.model[i]([torch.LongTensor(smiles_tensor).cuda(), 63 torch.LongTensor(length).cuda()], ---> 64 eval=True).detach().cpu().numpy()) 65 prediction = np.array(prediction).reshape(len(self.model), -1) 66 prediction = np.min(prediction, axis=0) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/models/Smiles2Label.py in forward(self, inp, eval) 41 else: 42 self.train() ---> 43 embedded = self.Embedding(inp) 44 output, _ = self.Encoder(embedded) 45 output = self.MLP(output) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/modules/embeddings/basic_embedding.py in forward(self, inp) 7 8 def forward(self, inp): ----> 9 embedded = self.embedding(inp) 10 return embedded /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/sparse.py in forward(self, input) 124 return F.embedding( 125 input, self.weight, self.padding_idx, self.max_norm, --> 126 self.norm_type, self.scale_grad_by_freq, self.sparse) 127 128 def extra_repr(self) -> str: /opt/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py in embedding(input, weight, padding_idx, max_norm, norm_type, scale_grad_by_freq, sparse) 1812 # remove once script supports set_grad_enabled 1813 _no_grad_embedding_renorm_(weight, input, max_norm, norm_type) -> 1814 return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse) 1815 1816 TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub < https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-253E&d=DwQFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=1UyQgH6hJ24o9H9c5HBPkIZTTgACDhz8kobR1mFM3yI&e=>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AAYPGLOG4EKA5TULTLDHX2DSLSQZDANCNFSM4SWVLIVA < https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AAYPGLOG4EKA5TULTLDHX2DSLSQZDANCNFSM4SWVLIVA-253E&d=DwQFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=cZfrAwvPGWJtxTjQXpNopDjE-6YRZWD6cyG59qU0y9U&e= .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-23issuecomment-2D712484495&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=gRSQ-wfSd5aALN-7K1HnwMMTzxf0A6KsZEqKnW0OxEw&e=>, or unsubscribe< https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AKBAWSFN5DXOXZWQJ4PDCADSLS7I5ANCNFSM4SWVLIVA&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=1Jpulkaf66vxDvRV9S4I6Dh3GvJgMBjbwZ5v5UzGI6U&e=

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/isayev/ReLeaSE/issues/38#issuecomment-712509158, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAYPGLOKQ2J4ASXV33PKTPDSLTIQTANCNFSM4SWVLIVA .

gmseabra commented 4 years ago

Right. What I wonder is that PyTorch is getting the data from OpenChem. Does that mean that OpenChem is in need of an update? Or would that be a localized thing with this notebook only?

From: Olexandr Isayev notifications@github.com Sent: Monday, October 19, 2020 8:15 PM To: isayev/ReLeaSE ReLeaSE@noreply.github.com Cc: de Miranda Seabra, Gustavo seabra@cop.ufl.edu; Author author@noreply.github.com Subject: Re: [isayev/ReLeaSE] LogP Example: "TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list" (#38)

[External Email]

Oh makes sense. Yeah, it does not work with the latest pytorch, you still have to run it with the old one.

On Mon, Oct 19, 2020 at 8:10 PM Gustavo Seabra notifications@github.com<mailto:notifications@github.com> wrote:

Well... I'm no wiz, but that much I had figured out.

The point is that there's no such thing as "my data": I'm just running the LogP Jupyter notebook from the git repo.

I assume it did work fine when it was created, with PyTorch 0.4. But maybe there was some change in PyTorch internals?

-- Gustavo Seabra

From: Olexandr Isayev notifications@github.com<mailto:notifications@github.com> Sent: Monday, October 19, 2020 6:51:58 PM To: isayev/ReLeaSE ReLeaSE@noreply.github.com<mailto:ReLeaSE@noreply.github.com> Cc: de Miranda Seabra, Gustavo seabra@cop.ufl.edu<mailto:seabra@cop.ufl.edu>; Author < author@noreply.github.commailto:author@noreply.github.com> Subject: Re: [isayev/ReLeaSE] LogP Example: "TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list" (#38)

[External Email]

Gustavo, I guess your data is a list, but the model expects a tensor.

On Mon, Oct 19, 2020 at 4:48 PM Gustavo Seabra notifications@github.com<mailto:notifications@github.com> wrote:

Hi,

I'm re-running the LogP example using current version of PyTorch, and the execution stops in the reinforcement loop due to a TypeError, as below. Are you aware of any changes in PyTorch that could be responsible for this? Is there a solution for it?

Thanks!

for i in range(n_iterations): for j in trange(n_policy, desc='Policy gradient...'): cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) rewards.append(simple_moving_average(rewards, cur_reward)) rl_losses.append(simple_moving_average(rl_losses, cur_loss))

plt.plot(rewards) plt.xlabel('Training iteration') plt.ylabel('Average reward') plt.show() plt.plot(rl_losses) plt.xlabel('Training iteration') plt.ylabel('Loss') plt.show()

smiles_cur, prediction_cur = estimate_and_update(RL_logp.generator, my_predictor, n_to_generate) print('Sample trajectories:') for sm in smiles_cur[:5]: print(sm)

with the error below:

Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]./release/data.py:98: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requiresgrad(True), rather than torch.tensor(sourceTensor). return torch.tensor(tensor).cuda() Policy gradient...: 0%| | 0/15 [00:00<?, ?it/s]

TypeError Traceback (most recent call last)

in 1 for i in range(n_iterations): 2 for j in trange(n_policy, desc='Policy gradient...'): ----> 3 cur_reward, cur_loss = RL_logp.policy_gradient(gen_data) 4 rewards.append(simple_moving_average(rewards, cur_reward)) 5 rl_losses.append(simple_moving_average(rl_losses, cur_loss)) ~/work/li/leadopt/generator/ReLeaSE/release/reinforcement.py in policy_gradient(self, data, n_batch, gamma, std_smiles, grad_clipping, **kwargs) 117 reward = self.get_reward(trajectory[1:-1], 118 self.predictor, --> 119 **kwargs) 120 121 # Converting string of characters into tensor in get_reward_logp(smiles, predictor, invalid_reward) 1 def get_reward_logp(smiles, predictor, invalid_reward=0.0): ----> 2 mol, prop, nan_smiles = predictor.predict([smiles]) 3 if len(nan_smiles) == 1: 4 return invalid_reward 5 if (prop[0] >= 1.0) and (prop[0] <= 4.0): ~/work/li/leadopt/generator/ReLeaSE/release/rnn_predictor.py in predict(self, smiles, use_tqdm) 62 self.model[i]([torch.LongTensor(smiles_tensor).cuda(), 63 torch.LongTensor(length).cuda()], ---> 64 eval=True).detach().cpu().numpy()) 65 prediction = np.array(prediction).reshape(len(self.model), -1) 66 prediction = np.min(prediction, axis=0) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/models/Smiles2Label.py in forward(self, inp, eval) 41 else: 42 self.train() ---> 43 embedded = self.Embedding(inp) 44 output, _ = self.Encoder(embedded) 45 output = self.MLP(output) /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), ~/work/source/repos/OpenChem/openchem/modules/embeddings/basic_embedding.py in forward(self, inp) 7 8 def forward(self, inp): ----> 9 embedded = self.embedding(inp) 10 return embedded /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs) 720 result = self._slow_forward(*input, **kwargs) 721 else: --> 722 result = self.forward(*input, **kwargs) 723 for hook in itertools.chain( 724 _global_forward_hooks.values(), /opt/anaconda3/lib/python3.7/site-packages/torch/nn/modules/sparse.py in forward(self, input) 124 return F.embedding( 125 input, self.weight, self.padding_idx, self.max_norm, --> 126 self.norm_type, self.scale_grad_by_freq, self.sparse) 127 128 def extra_repr(self) -> str: /opt/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py in embedding(input, weight, padding_idx, max_norm, norm_type, scale_grad_by_freq, sparse) 1812 # remove once script supports set_grad_enabled 1813 _no_grad_embedding_renorm_(weight, input, max_norm, norm_type) -> 1814 return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse) 1815 1816 TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not list — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub < https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-253E&d=DwQFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=1UyQgH6hJ24o9H9c5HBPkIZTTgACDhz8kobR1mFM3yI&e=%3E, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AAYPGLOG4EKA5TULTLDHX2DSLSQZDANCNFSM4SWVLIVA < https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AAYPGLOG4EKA5TULTLDHX2DSLSQZDANCNFSM4SWVLIVA-253E&d=DwQFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=cZfrAwvPGWJtxTjQXpNopDjE-6YRZWD6cyG59qU0y9U&e= .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-23issuecomment-2D712484495&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=gRSQ-wfSd5aALN-7K1HnwMMTzxf0A6KsZEqKnW0OxEw&e=%3E, or unsubscribe< https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AKBAWSFN5DXOXZWQJ4PDCADSLS7I5ANCNFSM4SWVLIVA&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=qtzQIvS0xBvZarMSMLncjHFM7UrivjdozY2-OFl0ji0&s=1Jpulkaf66vxDvRV9S4I6Dh3GvJgMBjbwZ5v5UzGI6U&e=

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/isayev/ReLeaSE/issues/38#issuecomment-712509158https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-23issuecomment-2D712509158-253E&d=DwQFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=g8Hs519lg9jWj_p-F-FnsJI-EhmmgcXTjLGNQuv1RKo&s=6SPqOUuzrRln09jU-VcZpfWl5GubhVc2FXTlcoQz21k&e=, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAYPGLOKQ2J4ASXV33PKTPDSLTIQTANCNFSM4SWVLIVAhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AAYPGLOKQ2J4ASXV33PKTPDSLTIQTANCNFSM4SWVLIVA-253E&d=DwQFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=g8Hs519lg9jWj_p-F-FnsJI-EhmmgcXTjLGNQuv1RKo&s=MYsNqa3KwF_-fv9I1b7PQLa7kcI7wKbdquVhY_iRdsE&e= .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_isayev_ReLeaSE_issues_38-23issuecomment-2D712510646&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=g8Hs519lg9jWj_p-F-FnsJI-EhmmgcXTjLGNQuv1RKo&s=HT1MLTUugX7Kn0T0bgyrpXos9S3Va9e908gYghirc70&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AKBAWSAWWKQGES6QIFEEBATSLTJB7ANCNFSM4SWVLIVA&d=DwMFaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=FJInzorQfnL2d-jllz5Qyw&m=g8Hs519lg9jWj_p-F-FnsJI-EhmmgcXTjLGNQuv1RKo&s=Oqb5l_M5tFPV8RqaubhB7sY6e3eSxKIFlsSnKT7MXqs&e=.