Closed jc-brandeis closed 2 years ago
errors while running "fit-the-data" using "COSMOS-HMM" model (tapqir v1.1.3). I did not change any of the parameters in tapqir.
/home/jchung/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/torch/distributions/gamma.py:71: UserWarning: Specified kernel cache directory could not be created! This disables kernel caching. Specified directory is /home/jchung/.cache/torch/kernels. This warning will appear only once per process. (Triggered internally at ../aten/src/ATen/native/cuda/jit_utils.cpp:860.) self.rate * value - torch.lgamma(self.concentration)) --------------------------------------------------------------------------- TypeError Traceback (most recent call last) File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/tapqir/gui.py:502, in fitCmd(b, layout, out, DEFAULTS) 500 DEFAULTS["priors"].update(layout["priors"].children[0].kwargs) 501 with out: --> 502 fit( 503 **layout.kwargs, 504 k_max=2, 505 funsor=False, 506 pykeops=True, 507 no_input=True, 508 progress_bar=tqdm_notebook, 509 ) 511 out.clear_output(wait=True) File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/tapqir/main.py:462, in fit(model, cuda, nbatch_size, fbatch_size, learning_rate, num_iter, k_max, matlab, funsor, pykeops, overwrite, no_input, progress_bar) 460 model.init(learning_rate, nbatch_size, fbatch_size) 461 try: --> 462 model.run(num_iter, progress_bar=progress_bar) 463 except CudaOutOfMemoryError: 464 logger.exception("Failed to fit the data") File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/tapqir/models/model.py:220, in Model.run(self, num_iter, progress_bar) 218 for i in progress_bar(range(num_iter)): 219 try: --> 220 self.iter_loss = self.svi.step() 221 # save a checkpoint every 200 iterations 222 if not self.iter % 200: File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/pyro/infer/svi.py:145, in SVI.step(self, *args, **kwargs) 143 # get loss and compute gradients 144 with poutine.trace(param_only=True) as param_capture: --> 145 loss = self.loss_and_grads(self.model, self.guide, *args, **kwargs) 147 params = set( 148 site["value"].unconstrained() for site in param_capture.trace.nodes.values() 149 ) 151 # actually perform gradient steps 152 # torch.optim objects gets instantiated for any params that haven't been seen yet File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/pyro/contrib/funsor/infer/elbo.py:20, in ELBO.loss_and_grads(self, model, guide, *args, **kwargs) 19 def loss_and_grads(self, model, guide, *args, **kwargs): ---> 20 loss = self.differentiable_loss(model, guide, *args, **kwargs) 21 loss.backward() 22 return loss.item() File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/pyro/contrib/funsor/infer/traceenum_elbo.py:104, in TraceMarkovEnum_ELBO.differentiable_loss(self, model, guide, *args, **kwargs) 94 def differentiable_loss(self, model, guide, *args, **kwargs): 95 96 # get batched, enumerated, to_funsor-ed traces from the guide and model 97 with plate( 98 size=self.num_particles 99 ) if self.num_particles > 1 else contextlib.ExitStack(), enum( (...) 102 else None 103 ): --> 104 guide_tr = trace(guide).get_trace(*args, **kwargs) 105 model_tr = trace(replay(model, trace=guide_tr)).get_trace(*args, **kwargs) 107 # extract from traces all metadata that we will need to compute the elbo File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/pyro/poutine/trace_messenger.py:198, in TraceHandler.get_trace(self, *args, **kwargs) 190 def get_trace(self, *args, **kwargs): 191 """ 192 :returns: data structure 193 :rtype: pyro.poutine.Trace (...) 196 Calls this poutine and returns its trace instead of the function's return value. 197 """ --> 198 self(*args, **kwargs) 199 return self.msngr.get_trace() File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/pyro/poutine/trace_messenger.py:174, in TraceHandler.__call__(self, *args, **kwargs) 170 self.msngr.trace.add_node( 171 "_INPUT", name="_INPUT", type="args", args=args, kwargs=kwargs 172 ) 173 try: --> 174 ret = self.fn(*args, **kwargs) 175 except (ValueError, RuntimeError) as e: 176 exc_type, exc_value, traceback = sys.exc_info() File ~/anaconda3/envs/tapqir-env/lib/python3.8/site-packages/tapqir/models/hmm.py:336, in hmm.guide(self) 334 for fdx in frames: 335 if self.vectorized: --> 336 fsx, fdx = fdx 337 fdx = torch.as_tensor(fdx) 338 fdx = fdx.unsqueeze(-1) TypeError: cannot unpack non-iterable int object
errors while running "fit-the-data" using "COSMOS-HMM" model (tapqir v1.1.3). I did not change any of the parameters in tapqir.