Training in progress, step 364

config.json

@@ -119,6 +119,7 @@
       "type": "audio"
     }
   },
+  "do_sample": true,
   "downsample_rate": 5,
   "dtype": "bfloat16",
   "encoder": {
@@ -276,6 +277,7 @@
   "repetition_penalty": 1.0,
   "router_aux_loss_coef": 0.01,
   "system_prompt": "",
+  "temperature": 0.7,
   "text_config": {
     "_name_or_path": "Qwen/Qwen3-1.7B",
     "architectures": [
@@ -341,6 +343,8 @@
   },
   "text_model_id": "Qwen/Qwen3-1.7B",
   "time_mask_length": 100,
+  "top_k": null,
+  "top_p": null,
   "transformers_version": "5.0.0.dev0",
   "use_cache": false,
   "use_lora": false,

generation_config.json

@@ -1,6 +1,6 @@
 {
   "bos_token_id": 151643,
-  "do_sample": false,
+  "do_sample": true,
   "eos_token_id": [
     151645,
     151643
@@ -12,6 +12,7 @@
   "num_beams": 1,
   "pad_token_id": 151643,
   "repetition_penalty": 1.0,
+  "temperature": 0.7,
   "transformers_version": "5.0.0.dev0",
   "use_cache": true
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:163c507b56adc124c364663e7b5561bcb16d9240ee9417d824edd335f9cd1623
+oid sha256:b8d27e4a7c907ced9bf93828c060f517ab52e14aa1d2d507a1ce23f8ae3f9435
 size 58732960