End of training

README.md

@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Whisper Large v3 Trained on Hindi
 
-This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the Common Voice 17.0 dataset.
+This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the Custom Hindi dataset dataset.
 
 ## Model description
 
@@ -37,17 +37,24 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 1e-05
-- train_batch_size: 4
-- eval_batch_size: 64
+- train_batch_size: 1
+- eval_batch_size: 16
 - seed: 42
+- distributed_type: multi-GPU
+- num_devices: 4
 - gradient_accumulation_steps: 16
 - total_train_batch_size: 64
+- total_eval_batch_size: 64
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
-- training_steps: 2500
+- lr_scheduler_warmup_steps: 250
+- num_epochs: 3.0
 - mixed_precision_training: Native AMP
 
+### Training results
+
+
+
 ### Framework versions
 
 - Transformers 4.41.1

config.json

@@ -25,6 +25,7 @@
   "encoder_layerdrop": 0.0,
   "encoder_layers": 32,
   "eos_token_id": 50257,
+  "forced_decoder_ids": null,
   "init_std": 0.02,
   "is_encoder_decoder": true,
   "mask_feature_length": 10,
@@ -42,7 +43,8 @@
   "num_mel_bins": 128,
   "pad_token_id": 50256,
   "scale_embedding": false,
-  "torch_dtype": "float32",
+  "suppress_tokens": [],
+  "torch_dtype": "float16",
   "transformers_version": "4.41.1",
   "use_cache": true,
   "use_weighted_layer_sum": false,

generation_config.json

@@ -48,6 +48,16 @@
   "bos_token_id": 50257,
   "decoder_start_token_id": 50258,
   "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      null
+    ],
+    [
+      2,
+      50360
+    ]
+  ],
   "is_multilingual": true,
   "lang_to_id": {
     "<|af|>": 50327,
@@ -151,7 +161,6 @@
     "<|yue|>": 50358,
     "<|zh|>": 50260
   },
-  "language": "hindi",
   "max_initial_timestamp_index": 50,
   "max_length": 448,
   "no_timestamps_token_id": 50364,
@@ -248,7 +257,6 @@
     50362,
     50363
   ],
-  "task": "transcribe",
   "task_to_id": {
     "transcribe": 50360,
     "translate": 50359

runs/May30_15-11-46_bhrathgpt-v1/events.out.tfevents.1717081927.bhrathgpt-v1.1780973.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3b92d362da1617a0bb339fc669110c4dd0fafca7d7523ec2733eeae34b5abe6
+size 6013

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c16c49102d4d640aad5c87fd2ada677c9a727a75db44136bf2fca0c34d6f3240
-size 4783
+oid sha256:3808877e186e53d02403c6cdca7d73cd3a22cc5292819e042fa3faa5156d51e0
+size 5615