allow float16 and flake8

theano/gpuarray/opt.py

@@ -1175,7 +1175,7 @@ def local_gpua_gemv(op, context_name, inputs, outputs):
 @op_lifter([tensor.blas.Gemm])
 @register_opt2([tensor.blas.Gemm], 'fast_compile')
 def local_gpua_gemm(op, context_name, inputs, outputs):
-    if inputs[0].dtype not in ['float32', 'float64']:
+    if inputs[0].dtype not in ['float16', 'float32', 'float64']:
         return
     if op.inplace:
         return gpugemm_inplace

theano/gpuarray/tests/test_basic_ops.py

@@ -26,7 +26,6 @@ from ..subtensor import GpuSubtensor
 
 from .config import mode_with_gpu, mode_without_gpu, test_ctx_name
 
-import pygpu
 from pygpu import gpuarray
 
 utt.seed_rng()
@@ -236,7 +235,7 @@ def gpu_alloc_expected(x, *shp):
 GpuAllocTester = makeTester(
     name="GpuAllocTester",
     # The +1 is there to allow the lift to the GPU.
-    op=lambda *args: alloc(*args)+1,
+    op=lambda *args: alloc(*args) + 1,
     gpu_op=GpuAlloc(test_ctx_name),
     cases=dict(
         correct01=(rand(), np.int32(7)),