增加qwen3vl

OpenGVLab/InternVL2_5-1B/bert4torch_config.json

@@ -43,7 +43,6 @@
     "vocab_size": 151674,
     "segment_vocab_size": 0,
     "rope_rank": "updown",
-    "max_position": 32768,
     "generation_config": {
         "tokenizer_config": {
             "allowed_special": [

Qwen/Qwen-1_8B-Chat/bert4torch_config.json

@@ -28,7 +28,7 @@
     "segment_vocab_size": 0,
     "skip_init": true,
     "rope_rank": "updown",
-    "max_position": 8192,
+    "max_position_embeddings": 8192,
     "generation_config": {
         "tokenizer_config": {
             "allowed_special": [

Qwen/Qwen-1_8B/bert4torch_config.json

@@ -29,7 +29,7 @@
     "segment_vocab_size": 0,
     "skip_init": true,
     "rope_rank": "updown",
-    "max_position": 8192,
+    "max_position_embeddings": 8192,
     "generation_config": {
         "tokenizer_config": {
             "allowed_special": [

Qwen/Qwen-7B-Chat/bert4torch_config.json

@@ -28,7 +28,7 @@
     "segment_vocab_size": 0,
     "skip_init": true,
     "rope_rank": "updown",
-    "max_position": 8192,
+    "max_position_embeddings": 8192,
     "generation_config": {
         "tokenizer_config": {
             "allowed_special": [

Qwen/Qwen-7B/bert4torch_config.json

@@ -29,7 +29,7 @@
     "segment_vocab_size": 0,
     "skip_init": true,
     "rope_rank": "updown",
-    "max_position": 8192,
+    "max_position_embeddings": 8192,
     "generation_config": {
         "tokenizer_config": {
             "allowed_special": [

Qwen/Qwen2.5-VL-32B-Instruct/bert4torch_config.json

@@ -0,0 +1,73 @@
+{
+    "model_type": "qwen2_5_vl",
+    "pos_emb_type": "rotary", 
+    "use_bias": false, 
+    "attention_bias": true,
+    "layer_norm_mode": "rmsnorm", 
+    "final_layernorm": true, 
+    "pre_layernorm": true,
+    "mlp_type": "LlamaFeedForward",
+    "template": "qwen2_vl",
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "eos_token_id": 151645,
+    "pad_token_id": 151643,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "hidden_act": "silu",
+    "hidden_size": 5120,
+    "initializer_range": 0.02,
+    "intermediate_size": 27648,
+    "max_position_embeddings": 128000,
+    "max_window_layers": 64,
+    "num_attention_heads": 40,
+    "num_hidden_layers": 64,
+    "num_key_value_heads": 8,
+    "layer_norm_eps": 1e-06,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "tie_word_embeddings": false,
+    "torch_dtype": "bfloat16",
+    "_attn_implementation": "sdpa",
+    "use_sliding_window": false,
+    "skip_init": true,
+    "segment_vocab_size": 0,
+    "rope_rank": "updown",
+    "convert_logits_dtype": "float32",
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "repetition_penalty": 1.05,
+        "temperature": 0.000001
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "hidden_size": 1280,
+        "in_chans": 3,
+        "intermediate_size": 3456,
+        "model_type": "qwen2_5_vl",
+        "out_hidden_size": 5120,
+        "spatial_patch_size": 14,
+        "tokens_per_second": 2,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation_internal": null
+    },
+    "rope_scaling": {
+        "mrope_section": [
+        16,
+        24,
+        24
+        ],
+        "rope_type": "default",
+        "type": "default"
+    },
+    "vocab_size": 152064
+}

Qwen/Qwen2.5-VL-3B-Instruct/bert4torch_config.json

@@ -0,0 +1,81 @@
+{
+    "model_type": "qwen2_5_vl",
+    "pos_emb_type": "rotary", 
+    "use_bias": false, 
+    "attention_bias": true,
+    "layer_norm_mode": "rmsnorm", 
+    "final_layernorm": true, 
+    "pre_layernorm": true,
+    "mlp_type": "LlamaFeedForward",
+    "template": "qwen2_vl",
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "eos_token_id": 151645,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "hidden_act": "silu",
+    "hidden_size": 2048,
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "max_position_embeddings": 128000,
+    "max_window_layers": 70,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 2,
+    "layer_norm_eps": 1e-06,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "_attn_implementation": "sdpa",
+    "use_sliding_window": false,
+    "skip_init": true,
+    "segment_vocab_size": 0,
+    "rope_rank": "updown",
+    "convert_logits_dtype": "float32",
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "repetition_penalty": 1.05,
+        "temperature": 0.000001
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "depth": 32,
+        "hidden_act": "silu",
+        "hidden_size": 1280,
+        "intermediate_size": 3420,
+        "num_heads": 16,
+        "in_chans": 3,
+        "out_hidden_size": 2048,
+        "patch_size": 14,
+        "spatial_merge_size": 2,
+        "spatial_patch_size": 14,
+        "window_size": 112,
+        "fullatt_block_indexes": [
+        7,
+        15,
+        23,
+        31
+        ],
+        "tokens_per_second": 2,
+        "_attn_implementation_internal": null
+    },
+    "rope_scaling": {
+        "type": "mrope",
+        "mrope_section": [
+            16,
+            24,
+            24
+        ]
+    },
+    "vocab_size": 151936
+}

Qwen/Qwen2.5-VL-7B-Instruct/bert4torch_config.json

@@ -0,0 +1,81 @@
+{
+    "model_type": "qwen2_5_vl",
+    "pos_emb_type": "rotary", 
+    "use_bias": false, 
+    "attention_bias": true,
+    "layer_norm_mode": "rmsnorm", 
+    "final_layernorm": true, 
+    "pre_layernorm": true,
+    "mlp_type": "LlamaFeedForward",
+    "template": "qwen2_vl",
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "eos_token_id": 151645,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "hidden_act": "silu",
+    "hidden_size": 3584,
+    "initializer_range": 0.02,
+    "intermediate_size": 18944,
+    "max_position_embeddings": 128000,
+    "max_window_layers": 28,
+    "num_attention_heads": 28,
+    "num_hidden_layers": 28,
+    "num_key_value_heads": 4,
+    "layer_norm_eps": 1e-06,
+    "rope_theta": 1000000.0,
+    "sliding_window": 32768,
+    "tie_word_embeddings": false,
+    "torch_dtype": "bfloat16",
+    "_attn_implementation": "sdpa",
+    "use_sliding_window": false,
+    "skip_init": true,
+    "segment_vocab_size": 0,
+    "rope_rank": "updown",
+    "convert_logits_dtype": "float32",
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "repetition_penalty": 1.05,
+        "temperature": 0.000001
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "depth": 32,
+        "hidden_act": "silu",
+        "hidden_size": 1280,
+        "intermediate_size": 3420,
+        "num_heads": 16,
+        "in_chans": 3,
+        "out_hidden_size": 3584,
+        "patch_size": 14,
+        "spatial_merge_size": 2,
+        "spatial_patch_size": 14,
+        "window_size": 112,
+        "fullatt_block_indexes": [
+        7,
+        15,
+        23,
+        31
+        ],
+        "tokens_per_second": 2,
+        "_attn_implementation_internal": null
+    },
+    "rope_scaling": {
+        "type": "mrope",
+        "mrope_section": [
+            16,
+            24,
+            24
+        ]
+    },
+    "vocab_size": 152064
+}

Qwen/Qwen3-VL-2B-Instruct/bert4torch_config.json

@@ -0,0 +1,88 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 2048,
+        "initializer_range": 0.02,
+        "intermediate_size": 6144,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 16,
+        "num_hidden_layers": 28,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": true,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "deepstack_visual_indexes": [
+            5,
+            11,
+            17
+        ],
+        "depth": 24,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1024,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4096,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 2048,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2,
+        "_attn_implementation_internal": null
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-2B-Thinking/bert4torch_config.json

@@ -0,0 +1,88 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 2048,
+        "initializer_range": 0.02,
+        "intermediate_size": 6144,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 16,
+        "num_hidden_layers": 28,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": true,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "deepstack_visual_indexes": [
+            5,
+            11,
+            17
+        ],
+        "depth": 24,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1024,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4096,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 2048,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2,
+        "_attn_implementation_internal": null
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-32B-Instruct/bert4torch_config.json

@@ -0,0 +1,87 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 5120,
+        "initializer_range": 0.02,
+        "intermediate_size": 25600,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 64,
+        "num_hidden_layers": 64,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": false,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+    "deepstack_visual_indexes": [
+        8,
+        16,
+        24
+        ],
+        "depth": 27,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1152,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4304,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 5120,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-32B-Thinking/bert4torch_config.json

@@ -0,0 +1,87 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 5120,
+        "initializer_range": 0.02,
+        "intermediate_size": 25600,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 64,
+        "num_hidden_layers": 64,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": false,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+    "deepstack_visual_indexes": [
+        8,
+        16,
+        24
+        ],
+        "depth": 27,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1152,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4304,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 5120,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-4B-Instruct/bert4torch_config.json

@@ -0,0 +1,88 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 2560,
+        "initializer_range": 0.02,
+        "intermediate_size": 9728,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 32,
+        "num_hidden_layers": 36,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": true,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "deepstack_visual_indexes": [
+            5,
+            11,
+            17
+        ],
+        "depth": 24,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1024,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4096,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 2560,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2,
+        "_attn_implementation_internal": null
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-4B-Thinking/bert4torch_config.json

@@ -0,0 +1,88 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 2560,
+        "initializer_range": 0.02,
+        "intermediate_size": 9728,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 32,
+        "num_hidden_layers": 36,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": true,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "deepstack_visual_indexes": [
+            5,
+            11,
+            17
+        ],
+        "depth": 24,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1024,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4096,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 2560,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2,
+        "_attn_implementation_internal": null
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-8B-Instruct/bert4torch_config.json

@@ -0,0 +1,87 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 4096,
+        "initializer_range": 0.02,
+        "intermediate_size": 12288,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 32,
+        "num_hidden_layers": 36,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": false,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "deepstack_visual_indexes": [
+        8,
+        16,
+        24
+        ],
+        "depth": 27,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1152,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4304,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 4096,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

Qwen/Qwen3-VL-8B-Thinking/bert4torch_config.json

@@ -0,0 +1,87 @@
+{
+    "model_type": "qwen3_vl",
+    "template": "qwen3_vl",
+    "torch_dtype": "bfloat16",
+    "skip_init": true,
+    "image_token_id": 151655,
+    "video_token_id": 151656,
+    "text_config": {
+        "pos_emb_type": "rotary",
+        "use_bias": false,
+        "layer_norm_mode": "rmsnorm",
+        "final_layernorm": true,
+        "pre_layernorm": true,
+        "mlp_type": "LlamaFeedForward",
+        "attn_type": "Qwen3Attention",
+        "attention_dropout": 0.0,
+        "bos_token_id": 151643,
+        "eos_token_id": 151645,
+        "image_token_id": 151655,
+        "video_token_id": 151656,
+        "hidden_act": "silu",
+        "hidden_size": 4096,
+        "initializer_range": 0.02,
+        "intermediate_size": 12288,
+        "max_position_embeddings": 262144,
+        "num_attention_heads": 32,
+        "num_hidden_layers": 36,
+        "num_key_value_heads": 8,
+        "layer_norm_eps": 1e-06,
+        "rope_theta": 5000000,
+        "tie_word_embeddings": false,
+        "torch_dtype": "bfloat16",
+        "_attn_implementation": "sdpa",
+        "use_sliding_window": false,
+        "skip_init": true,
+        "segment_vocab_size": 0,
+        "rope_rank": "updown",
+        "convert_logits_dtype": "float32",
+        "rope_scaling": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                24,
+                20,
+                20
+            ],
+            "rope_type": "mrope_interleaved"
+        },
+        "vocab_size": 151936
+    },
+    "vision_start_token_id": 151652,
+    "vision_end_token_id": 151653,
+    "vision_token_id": 151654,
+    "vision_config": {
+        "deepstack_visual_indexes": [
+        8,
+        16,
+        24
+        ],
+        "depth": 27,
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1152,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4304,
+        "model_type": "qwen3_vl",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 4096,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2
+    },
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true
+        },
+        "eos_token_id": [
+            151643,
+            151645
+        ],
+        "max_length": 32768,
+        "top_p": 0.8,
+        "top_k": 20,
+        "temperature": 0.7,
+        "repetition_penalty": 1.0
+    }
+}

meta-llama/Llama-3.2-11B-Vision-Instruct/bert4torch_config.json

@@ -35,7 +35,6 @@
         "initializer_range": 0.02,
         "intermediate_size": 14336,
         "max_position_embeddings": 131072,
-        "max_position": 131072,
         "num_attention_heads": 32,
         "num_hidden_layers": 40,
         "num_key_value_heads": 8,

meta-llama/Llama-3.2-11B-Vision/bert4torch_config.json

@@ -35,7 +35,6 @@
         "initializer_range": 0.02,
         "intermediate_size": 14336,
         "max_position_embeddings": 131072,
-        "max_position": 131072,
         "num_attention_heads": 32,
         "num_hidden_layers": 40,
         "num_key_value_heads": 8,