softmax: Correct to include exponential

sw/dnn/src/softmax.h

@@ -1,12 +1,13 @@
 // Copyright 2020 ETH Zurich and University of Bologna.
 // Licensed under the Apache License, Version 2.0, see LICENSE for details.
 // SPDX-License-Identifier: Apache-2.0
+//
+// Viviane Potocnik, ETH Zurich, <[email protected]>
 
 #pragma once
 
 #include "math.h"
 #include "snrt.h"
-// #include "printf.h"
 #include "utils.h"
 
 /**
@@ -41,6 +42,10 @@ typedef struct softmax_layer_struct {
     precision_t dtype;
 } softmax_layer_t;
 
+/* Dump function for single-cycle printf. */
+dump_float(input, 7);
+dump_float(output, 8);
+
 /**
  * Implementation of the SoftMax layer.
  */
@@ -50,15 +55,19 @@ static inline void softmax_fp32(float *input, float *output, int32_t ldI,
     float max_core = 0.0;  // max value of the current core
     float sum = 0.0;       // sum of the exp values of the current core
 
-    // uint32_t compute_id = snrt_global_core_idx();
-    // uint32_t num_cores = snrt_cluster_compute_core_num();
+    uint32_t compute_id = snrt_global_core_idx();
+    uint32_t num_cores = snrt_cluster_compute_core_num();
+
+    // printf("Hello from core [%d/%d]\n", compute_id, num_cores);
+    // dump_softmax(compute_id);
 
     for (int32_t b = 0; b < batch_size; b++) {
         for (int32_t s = 0; s < seq_len; s++) {
             max_core = -INFINITY;
             sum = 0.0;
 
             for (int32_t i = 0; i < input_samples; i++) {
+                // dump_input(input[b * batch_offset + s * ldI + i]);
                 if (input[b * batch_offset + s * ldI + i] > max_core) {
                     max_core = input[b * batch_offset + s * ldI + i];
                 }
@@ -69,18 +78,15 @@ static inline void softmax_fp32(float *input, float *output, int32_t ldI,
                 output[b * batch_offset + s * ldI + i] =
                     // FIXME: Below code is erroring due to the standard math
                     // lib conflict
-                    // TODO: Try out with musl lib
-                    // expf(input[b * batch_offset + s * ldI + i] - max_core);
-                    // FIXME: actually there should be an exponentiation
-                    input[b * batch_offset + s * ldI + i] - max_core;
+                    expf(input[b * batch_offset + s * ldI + i] - max_core);
+                // input[b * batch_offset + s * ldI + i] - max_core;
                 sum += output[b * batch_offset + s * ldI + i];
             }
 
             // compute the softmax value of the current row
             for (int32_t i = 0; i < input_samples; i++) {
-                // INFO: DIVSQRT unit MUST be activated in the cluster
-                // configuration
                 output[b * batch_offset + s * ldI + i] /= sum;
+                // dump_output(output[b * batch_offset + s * ldI + i]);
                 // printf("output[%d] = %f\n", compute_id * input_samples + b *
                 // batch_offset + s * ldI + i,
                 //        output[b * batch_offset + s * ldI + i]);