diff --git a/ggml/src/ggml-cuda/im2col.cu b/ggml/src/ggml-cuda/im2col.cu
index 3d0d8d4e..97fe6192 100644
--- a/ggml/src/ggml-cuda/im2col.cu
+++ b/ggml/src/ggml-cuda/im2col.cu
@@ -69,7 +69,7 @@ void ggml_cuda_op_im2col(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
     float * dst_d = (float *)dst->data;
     cudaStream_t stream = ctx.stream();
 
-    GGML_ASSERT(src0->type == GGML_TYPE_F16);
+    //GGML_ASSERT(src0->type == GGML_TYPE_F16);
     GGML_ASSERT(src1->type == GGML_TYPE_F32);
     GGML_ASSERT( dst->type == GGML_TYPE_F16 || dst->type == GGML_TYPE_F32);
 
diff --git a/ggml/src/ggml.c b/ggml/src/ggml.c
index 96e411d3..81077f5f 100644
--- a/ggml/src/ggml.c
+++ b/ggml/src/ggml.c
@@ -18805,7 +18805,6 @@ static void ggml_compute_forward_im2col_f16(
     const struct ggml_tensor * src0 = dst->src[0];
     const struct ggml_tensor * src1 = dst->src[1];
 
-    GGML_ASSERT(src0->type == GGML_TYPE_F16);
     GGML_ASSERT(src1->type == GGML_TYPE_F32);
     GGML_ASSERT( dst->type == GGML_TYPE_F16);
 
@@ -18836,7 +18835,6 @@ static void ggml_compute_forward_im2col_f16(
     int ofs0 = is_2D ? nb13 : nb12;
     int ofs1 = is_2D ? nb12 : nb11;
 
-    GGML_ASSERT(nb00 == sizeof(ggml_fp16_t));
     GGML_ASSERT(nb10 == sizeof(float));
 
     // im2col: [N, IC, IH, IW] => [N, OH, OW, IC*KH*KW]