初始化项目，由ModelHub XC社区提供模型

Model: ibm-granite/granite-speech-3.3-8b Source: Original Platform
2026-10-29 07:25:18 +08:00
commit ffada56896
26 changed files with 295842 additions and 0 deletions
--- a/config.json
+++ b/config.json
@@ -0,0 +1,79 @@
+{
+    "architectures": [
+        "GraniteSpeechForConditionalGeneration"
+    ],
+    "audio_token_index": 49159,
+    "downsample_rate": 5,
+    "encoder_config": {
+        "context_size": 200,
+        "conv_expansion_factor": 2,
+        "conv_kernel_size": 15,
+        "dim_head": 128,
+        "dropout": 0.1,
+        "feedforward_mult": 4,
+        "hidden_dim": 1024,
+        "input_dim": 160,
+        "max_pos_emb": 512,
+        "model_type": "granite_speech_encoder",
+        "num_heads": 8,
+        "num_layers": 16,
+        "output_dim": 256
+    },
+    "has_lora_adapter": true,
+    "initializer_range": 0.02,
+    "model_type": "granite_speech",
+    "projector_config": {
+        "_attn_implementation_autoset": true,
+        "attention_probs_dropout_prob": 0.1,
+        "cross_attention_frequency": 1,
+        "encoder_hidden_size": 1024,
+        "hidden_act": "gelu",
+        "hidden_dropout_prob": 0.1,
+        "hidden_size": 1024,
+        "initializer_range": 0.02,
+        "intermediate_size": 4096,
+        "layer_norm_eps": 1e-12,
+        "max_position_embeddings": 2048,
+        "model_type": "blip_2_qformer",
+        "num_attention_heads": 16,
+        "num_hidden_layers": 2,
+        "position_embedding_type": "absolute",
+        "use_qformer_text_input": false,
+        "vocab_size": 30522
+    },
+    "text_config": {
+        "_name_or_path": "/proj/speech/data/granite-3.3-8b-instruct/r250409a",
+        "architectures": [
+            "GraniteForCausalLM"
+        ],
+        "attention_bias": false,
+        "attention_dropout": 0.0,
+        "attention_multiplier": 0.0078125,
+        "bos_token_id": 0,
+        "embedding_multiplier": 12.0,
+        "eos_token_id": 0,
+        "hidden_act": "silu",
+        "hidden_size": 4096,
+        "initializer_range": 0.02,
+        "intermediate_size": 12800,
+        "logits_scaling": 16.0,
+        "max_position_embeddings": 131072,
+        "mlp_bias": false,
+        "model_type": "granite",
+        "num_attention_heads": 32,
+        "num_hidden_layers": 40,
+        "num_key_value_heads": 8,
+        "pad_token_id": 0,
+        "residual_multiplier": 0.22,
+        "rms_norm_eps": 1e-05,
+        "rope_scaling": null,
+        "rope_theta": 10000000.0,
+        "tie_word_embeddings": true,
+        "torch_dtype": "bfloat16",
+        "use_cache": true,
+        "vocab_size": 49160
+    },
+    "torch_dtype": "bfloat16",
+    "transformers_version": "4.52.4",
+    "window_size": 15
+}