Merge pull request #50 from Mangio621/auto-epoch-log

TensorBoard logging is now per-steps in an epoch

infer-web.py

@@ -2,6 +2,7 @@ import os
 import shutil
 import sys
 import json  # Mangio fork using json for preset saving
+import math
 
 import signal
 
@@ -943,6 +944,23 @@ def change_f0(if_f0_3, sr2, version19):  # f0method8,pretrained_G14,pretrained_D
     )
 
 
+global log_interval
+
+
+def set_log_interval(exp_dir, batch_size12):
+    log_interval = 1
+
+    folder_path = os.path.join(exp_dir, "1_16k_wavs")
+
+    if os.path.exists(folder_path) and os.path.isdir(folder_path):
+        wav_files = [f for f in os.listdir(folder_path) if f.endswith(".wav")]
+        if wav_files:
+            sample_size = len(wav_files)
+            log_interval = math.ceil(sample_size / batch_size12)
+
+    return log_interval
+
+
 # but3.click(click_train,[exp_dir1,sr2,if_f0_3,save_epoch10,total_epoch11,batch_size12,if_save_latest13,pretrained_G14,pretrained_D15,gpus16])
 def click_train(
     exp_dir1,
@@ -969,6 +987,9 @@ def click_train(
         if version19 == "v1"
         else "%s/3_feature768" % (exp_dir)
     )
+
+    log_interval = set_log_interval(exp_dir, batch_size12)
+
     if if_f0_3:
         f0_dir = "%s/2a_f0" % (exp_dir)
         f0nsf_dir = "%s/2b-f0nsf" % (exp_dir)
@@ -1038,7 +1059,7 @@ def click_train(
         ####
         cmd = (
             config.python_cmd
-            + " train_nsf_sim_cache_sid_load_pretrain.py -e %s -sr %s -f0 %s -bs %s -g %s -te %s -se %s %s %s -l %s -c %s -sw %s -v %s"
+            + " train_nsf_sim_cache_sid_load_pretrain.py -e %s -sr %s -f0 %s -bs %s -g %s -te %s -se %s %s %s -l %s -c %s -sw %s -v %s -li %s"
             % (
                 exp_dir1,
                 sr2,
@@ -1053,12 +1074,13 @@ def click_train(
                 1 if if_cache_gpu17 == i18n("是") else 0,
                 1 if if_save_every_weights18 == i18n("是") else 0,
                 version19,
+                log_interval,
             )
         )
     else:
         cmd = (
             config.python_cmd
-            + " train_nsf_sim_cache_sid_load_pretrain.py -e %s -sr %s -f0 %s -bs %s -te %s -se %s %s %s -l %s -c %s -sw %s -v %s"
+            + " train_nsf_sim_cache_sid_load_pretrain.py -e %s -sr %s -f0 %s -bs %s -te %s -se %s %s %s -l %s -c %s -sw %s -v %s -li %s"
             % (
                 exp_dir1,
                 sr2,
@@ -1072,6 +1094,7 @@ def click_train(
                 1 if if_cache_gpu17 == i18n("是") else 0,
                 1 if if_save_every_weights18 == i18n("是") else 0,
                 version19,
+                log_interval,
             )
         )
     print(cmd)

train/utils.py

@@ -352,6 +352,9 @@ def get_hparams(init=True):
         required=True,
         help="if caching the dataset in GPU memory, 1 or 0",
     )
+    parser.add_argument(
+        "-li", "--log_interval", type=int, required=True, help="log interval"
+    )
 
     args = parser.parse_args()
     name = args.experiment_dir
@@ -391,6 +394,16 @@ def get_hparams(init=True):
     hparams.save_every_weights = args.save_every_weights
     hparams.if_cache_data_in_gpu = args.if_cache_data_in_gpu
     hparams.data.training_files = "%s/filelist.txt" % experiment_dir
+
+    hparams.train.log_interval = args.log_interval
+
+    # Update log_interval in the 'train' section of the config dictionary
+    config["train"]["log_interval"] = args.log_interval
+
+    # Save the updated config back to the config_save_path
+    with open(config_save_path, "w") as f:
+        json.dump(config, f, indent=4)
+
     return hparams