Upload quantized Marvis TTS 100M v0.2 with all configs and test samples

config.json

@@ -0,0 +1,17 @@
+{
+  "architectures": [
+    "CsmModel"
+  ],
+  "model_type": "csm",
+  "hidden_size": 1024,
+  "num_hidden_layers": 24,
+  "num_attention_heads": 16,
+  "intermediate_size": 2816,
+  "max_position_embeddings": 4096,
+  "torch_dtype": "float16",
+  "vocab_size": 200000,
+  "quantization": {
+    "method": "8-bit-bitsandbytes",
+    "load_in_8bit": true
+  }
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "do_sample": false,
+  "max_length": 20,
+  "num_beams": 1
+}

quantization_config.json

@@ -4,6 +4,13 @@
   "model_size_original_mb": 930,
   "model_size_quantized_mb": 465,
   "compression_ratio": "50%",
-  "dtype_original": "float16",
-  "dtype_quantized": "int8"
+  "inference_framework": "transformers",
+  "tested_samples": [
+    "Hello, this is a test of the quantized Marvis TTS model.",
+    "Marvis TTS provides efficient real-time text-to-speech synthesis.",
+    "The quantized model maintains high quality while reducing memory usage.",
+    "You can use this model for voice synthesis on edge devices."
+  ],
+  "quantization_date": "2025-11-10",
+  "all_samples_passed": true
 }

tokenizer_config.json

@@ -0,0 +1,6 @@
+{
+  "model_max_length": 1000000000000000019884200,
+  "padding_side": "right",
+  "tokenizer_class": "AutoTokenizer",
+  "truncation_side": "right"
+}