Training in progress, step 14000

config.json

@@ -64,11 +64,9 @@
       "type": "audio"
     }
   },
-  "decoder_model_name": "HuggingFaceTB/SmolLM3-3B",
   "downsample_rate": 16,
   "dtype": "bfloat16",
   "encoder_dim": 1280,
-  "encoder_model_name": "openai/whisper-large-v3-turbo",
   "inference_diversity_penalty": 0.0,
   "inference_warmup_tokens": 10,
   "label_smoothing": 0.0,
@@ -83,13 +81,12 @@
   "projector_dropout": 0.0,
   "projector_hidden_dim": null,
   "projector_init_std": 0.02,
-  "projector_input_noise": 0.02,
+  "projector_input_noise": 0.0,
   "projector_num_layers": 2,
   "projector_pool_stride": 2,
   "projector_type": "mlp",
   "router_aux_loss_coef": 0.01,
   "system_prompt": "/no_think /system_override",
-  "temperature": 0.1,
   "text_config": {
     "_name_or_path": "HuggingFaceTB/SmolLM3-3B",
     "architectures": [

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1530d120be6020e944be23373fcca73ec2a4dcac75f473117ad95a68332322fc
+oid sha256:25d44f082c613a5981137d73a0fa57706f3dd2d2fed56ad80f553b27b4f11471
 size 23462224

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a614500a8c2b6ea97976e7dbd2e49f64b5adef85d1fffcffceebe71b6fd4ce58
-size 5905
+oid sha256:316f75ee83a5b4c47f214306644d6417f18aaad67f6f41d0e1950ffdaafdc9f9
+size 5969