[ICCV'21] Official implementation for the paper Social NCE: Contrastive Learning of Socially-aware Motion Representations

Thanks for your wonderful repo! May I ask what does SAIL policy refer to? Based on the paper I only find SARL but not SAIL. Is SAIL an upgrade version of SARL? Thanks!

Hello,

first of all, thank you for the publicly available code! I have 2 understanding questions regarding the training:

1. as far as I understand you determine the positive and negative samples based on the ground truth at time t, with the horizon of e.g. 5 time steps. What about the possible collisions before and after the one point?
2. What is the ground truth in reinforcement learning? Do you use here a linear model?

Best regards

hello, thanks for you available code. i met error like this when i run "python imitate.py --contrast_weight=2.0 --contrast_sampling='local' --gpu" in crowd_nav path, the error information is: RuntimeError: CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc) it seems that the attention's linear operation got the wrong shapes, what should i do? thanks :)

Thanks for your hard work.

I have some questions about this code for alpha in alpha_list: sample_boundary.append(neg_seed * alpha + pos_seed.unsqueeze(2) * (1-alpha)) Looks like α* neg_ seeds + (1- α)* pos_ seeds What is the purpose of this code

from sampling.py

    # primary-neighbor boundary
    if self.num_boundary > 0:
        alpha_list = torch.linspace(self.ratio_boundary, 1.0, steps=self.num_boundary) #0.5-1 分成0-9份 参数决定
        sample_boundary = []
        for alpha in alpha_list:
            sample_boundary.append(neg_seed * alpha + pos_seed.unsqueeze(2) * (1-alpha))
        sample_boundary = torch.cat(sample_boundary, axis=2)
        sample_neg = torch.cat([sample_boundary, sample_territory], axis=2)
    else:
        sample_neg = sample_territory

Thanks!

Hi I am about to try your code on the Trajnet++.

I am trying to get the FDE score of 1.14 .

Did you train on the whole (with cff) training dataset? How many epochs? And with which parameters?

If I reach your performance I will delete the submission if desired.

Thanks in advance Many greetings

When I download data, pip install gdown && gdown https://drive.google.com/uc?id=1D2guAxD_EgrKnJFMcLSBkf10SOagz0mr, it occurs that: Traceback (most recent call last): File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connection.py", line 169, in _new_conn conn = connection.create_connection( File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/util/connection.py", line 96, in create_connection raise err File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/util/connection.py", line 86, in create_connection sock.connect(sa) OSError: [Errno 101] Network is unreachable

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connectionpool.py", line 699, in urlopen httplib_response = self._make_request( File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connectionpool.py", line 382, in _make_request self._validate_conn(conn) File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connectionpool.py", line 1010, in _validate_conn conn.connect() File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connection.py", line 353, in connect conn = self._new_conn() File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connection.py", line 181, in _new_conn raise NewConnectionError( urllib3.exceptions.NewConnectionError: <urllib3.connection.HTTPSConnection object at 0x7f96a64c1100>: Failed to establish a new connection: [Errno 101] Network is unreachable

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "/home/lyl/miniconda3/lib/python3.8/site-packages/requests/adapters.py", line 439, in send resp = conn.urlopen( File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/connectionpool.py", line 755, in urlopen retries = retries.increment( File "/home/lyl/miniconda3/lib/python3.8/site-packages/urllib3/util/retry.py", line 573, in increment raise MaxRetryError(_pool, url, error or ResponseError(cause)) urllib3.exceptions.MaxRetryError: HTTPSConnectionPool(host='drive.google.com', port=443): Max retries exceeded with url: /uc?id=1awXDsRQcmgacj7nUhPzwb5UMNZCJCjvu (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7f96a64c1100>: Failed to establish a new connection: [Errno 101] Network is unreachable'))

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "/home/lyl/miniconda3/bin/gdown", line 8, in sys.exit(main()) File "/home/lyl/miniconda3/lib/python3.8/site-packages/gdown/cli.py", line 95, in main download( File "/home/lyl/miniconda3/lib/python3.8/site-packages/gdown/download.py", line 77, in download res = sess.get(url, stream=True) File "/home/lyl/miniconda3/lib/python3.8/site-packages/requests/sessions.py", line 555, in get return self.request('GET', url, **kwargs) File "/home/lyl/miniconda3/lib/python3.8/site-packages/requests/sessions.py", line 542, in request resp = self.send(prep, **send_kwargs) File "/home/lyl/miniconda3/lib/python3.8/site-packages/requests/sessions.py", line 655, in send r = adapter.send(request, **kwargs) File "/home/lyl/miniconda3/lib/python3.8/site-packages/requests/adapters.py", line 516, in send raise ConnectionError(e, request=request) requests.exceptions.ConnectionError: HTTPSConnectionPool(host='drive.google.com', port=443): Max retries exceeded with url: /uc?id=1awXDsRQcmgacj7nUhPzwb5UMNZCJCjvu (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7f96a64c1100>: Failed to establish a new connection: [Errno 101] Network is unreachable'))

Can you fix it?Thanks.

Thank you for your hard work. in the file "crowd_nav/utils/demonstrate.py"

   from crowd_nav.utils.frames import FrameStack
  from crowd_nav.utils.trainer import Imitator
  from crowd_nav.utils.memory import ReplayMemory

I cannot find "Imitator" in crowd_nav.utils.trainer. will you help me ?

Hello, me and my team are trying to analyze your works(both of crowdnav and social-nce) as term project. While I run test.py I face two problems. 1st. In square scenario, there is too large decreasing of performance. WHY?

!python test.py --policy='sail' --square --model_file=data/output/imitate-event-data-0.50-weight-2.0-horizon-4-temperature-0.20-nboundary-0/policy_net.pth (skip) 2021-12-05 08:55:59, INFO: TEST success: 0.58, collision: 0.41, nav time: 10.43, reward: 0.1095 +- 0.2644 2021-12-05 08:55:59, INFO: Frequency of being in danger: 1.23

2nd. In mixed scenario, (I add argument in parser), error is occur Traceback (most recent call last):

File "test.py", line 129, in main() File "test.py", line 126, in main explorer.run_k_episodes(env.case_size[args.phase], args.phase, print_failure=False) File "/content/drive/My Drive/ColabFiles/robotvision/social-nce/crowd_nav/utils/explorer.py", line 60, in run_k_episodes action = self.robot.act(ob) File "/content/drive/My Drive/ColabFiles/robotvision/social-nce/crowd_sim/envs/utils/robot.py", line 13, in act action = self.policy.predict(state) File "/content/drive/My Drive/ColabFiles/robotvision/social-nce/crowd_nav/policy/sail.py", line 123, in predict action = self.model(self.last_state[0], self.last_state[1])[0].squeeze() File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py", line 1102, in _call_impl return forward_call(*input, **kwargs) File "/content/drive/My Drive/ColabFiles/robotvision/social-nce/crowd_nav/policy/sail.py", line 73, in forward human_state = self.transform.transform_frame(crowd_obsv) File "/content/drive/My Drive/ColabFiles/robotvision/social-nce/crowd_nav/utils/transform.py", line 14, in transform_frame state = torch.cat([frame, relative], axis=2) RuntimeError: Sizes of tensors must match except in dimension 2. Expected size 1 but got size 5 for tensor number 1 in the list.

If you already have knew about those problem, please share the answers. Thank you for reading