error:Shape mismatch: y has 2000 cols but z has 1 cols

[LinneLee]

when running the code, I got this error message:
Exception has occurred: ValueError (note: full exception trace is shown but execution is paused at: _run_module_as_main)
Shape mismatch: y has 2000 cols but z has 1 cols
Apply node that caused the error: Gemm{no_inplace}(Reshape{2}.0, TensorConstant{2.0}, <TensorType(float32, matrix)>, X.T, TensorConstant{-1.0})
Toposort index: 13
Inputs types: [TensorType(float32, matrix), TensorType(float32, scalar), TensorType(float32, matrix), TensorType(float32, matrix), TensorType(float32, scalar)]
Inputs shapes: [(2000, 1), (), (2000, 100), (100, 2000), ()]
Inputs strides: [(4, 4), (), (400, 4), (4, 400), ()]
Inputs values: ['not shown', array(2., dtype=float32), 'not shown', 'not shown', array(-1., dtype=float32)]
Outputs clients: [[Elemwise{Composite{exp((((-i0) + i1) / i2))}}[(0, 0)](Reshape{2}.0, Gemm{no_inplace}.0, Reshape{2}.0)]]

HINT: Re-running with most Theano optimization disabled could give you a back-trace of when this node was created. This can be done with by setting the Theano flag 'optimizer=fast_compile'. If that does not work, Theano optimizations can be disabled with 'optimizer=None'.
HINT: Use the Theano flag 'exception_verbosity=high' for a debugprint and storage map footprint of this apply node.
File "/home/anaconda3/envs/tp-gpd/lib/python3.8/site-packages/theano/compile/function_module.py", line 903, in call
self.fn() if output_subset is None else
ValueError: Shape mismatch: y has 2000 cols but z has 1 cols

During handling of the above exception, another exception occurred:

File "/home/anaconda3/envs/tp-gpd/lib/python3.8/site-packages/theano/compile/function_module.py", line 914, in call
gof.link.raise_with_op(
File "/home/anaconda3/envs/tp-gpd/lib/python3.8/site-packages/six.py", line 718, in reraise
raise value.with_traceback(tb)
File "/home/anaconda3/envs/tp-gpd/lib/python3.8/site-packages/theano/gof/link.py", line 325, in raise_with_op
reraise(exc_type, exc_value, exc_trace)
File "/home/anaconda3/envs/tp-gpd/lib/python3.8/site-packages/theano/compile/function_module.py", line 914, in call
gof.link.raise_with_op(
File "/home/work space/tp_gnn/group_detect/thesne/model/core.py", line 92, in find_sigma
e = update_intervals()
File "/home/work space/tp_gnn/group_detect/thesne/model/dynamic_tsne.py", line 205, in dynamic_tsne
find_sigma(X_shared, sigma_shared, N, perplexity, sigma_iters,
File "/home/work space/tp_gnn/group_detect/thesne/examples/gaussians.py", line 49, in main
Ys = dynamic_tsne(Xs, perplexity=70, lmbda=0.1, verbose=1, sigma_iters=50,
File "/home/work space/tp_gnn/group_detect/thesne/examples/gaussians.py", line 57, in
main()
File "/home/anaconda3/envs/tp-gpd/lib/python3.8/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/anaconda3/envs/tp-gpd/lib/python3.8/runpy.py", line 192, in _run_module_as_main (Current frame)
return _run_code(code, main_globals, None,
ValueError: Shape mismatch: y has 2000 cols but z has 1 cols
Apply node that caused the error: Gemm{no_inplace}(Reshape{2}.0, TensorConstant{2.0}, <TensorType(float32, matrix)>, X.T, TensorConstant{-1.0})
Toposort index: 13
Inputs types: [TensorType(float32, matrix), TensorType(float32, scalar), TensorType(float32, matrix), TensorType(float32, matrix), TensorType(float32, scalar)]
Inputs shapes: [(2000, 1), (), (2000, 100), (100, 2000), ()]
Inputs strides: [(4, 4), (), (400, 4), (4, 400), ()]
Inputs values: ['not shown', array(2., dtype=float32), 'not shown', 'not shown', array(-1., dtype=float32)]
Outputs clients: [[Elemwise{Composite{exp((((-i0) + i1) / i2))}}[(0, 0)](Reshape{2}.0, Gemm{no_inplace}.0, Reshape{2}.0)]]

HINT: Re-running with most Theano optimization disabled could give you a back-trace of when this node was created. This can be done with by setting the Theano flag 'optimizer=fast_compile'. If that does not work, Theano optimizations can be disabled with 'optimizer=None'.
HINT: Use the Theano flag 'exception_verbosity=high' for a debugprint and storage map footprint of this apply node.

The error vanishes when I sent only one data (X) with shape (1, -1) in the tsne function, but it seems the wrong way to resolve this since I still can not get correct output. Could you give me some hint to solve this error?