add PaddleOCR-VL

PaddlePaddle/PaddleOCR-VL/bert4torch_config.json

@@ -0,0 +1,77 @@
+{
+    "model_type": "paddleocr_vl",
+    "template": "paddleocr_vl",
+    "bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 0,
+    "eos_token_id": 2,
+    "image_token_id": 100295,
+    "video_token_id": 101307,
+    "hidden_act": "silu",
+    "hidden_size": 1024,
+    "initializer_range": 0.02,
+    "intermediate_size": 3072,
+    "max_position_embeddings": 131072,
+    "max_window_layers": 28,
+    "attention_head_size": 128,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 18,
+    "num_key_value_heads": 2,
+    "layer_norm_eps": 1e-05,
+    "rope_theta": 500000.0,
+    "sliding_window": null,
+    "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": 2,
+        "max_length": 32768,
+        "repetition_penalty": 1.0,
+        "temperature": 0.01,
+        "top_p": 0.001,
+        "top_k": 1
+    },
+    "vision_start_token_id": 101305,
+    "vision_config": {
+        "architectures": [
+        "SiglipVisionModel"
+        ],
+        "attention_dropout": 0.0,
+        "auto_map": {
+        "AutoConfig": "configuration_paddleocr_vl.PaddleOCRVLConfig",
+        "AutoModel": "modeling_paddleocr_vl.SiglipVisionModel"
+        },
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1152,
+        "image_size": 384,
+        "intermediate_size": 4304,
+        "layer_norm_eps": 1e-06,
+        "model_type": "paddleocr_vl",
+        "num_attention_heads": 16,
+        "num_channels": 3,
+        "num_hidden_layers": 27,
+        "pad_token_id": 0,
+        "patch_size": 14,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2,
+        "tokens_per_second": 2,
+        "torch_dtype": "bfloat16"
+    },
+    "rope_scaling": {
+        "type": "mrope",
+        "mrope_section": [
+            16,
+            24,
+            24
+        ]
+    },
+    "vocab_size": 103424
+}

baidu/ERNIE-4.5-0.3B-Base-PT/bert4torch_config.json

@@ -0,0 +1,36 @@
+{
+    "model": "ernie4_5",
+    "template": "pretrained_text_continuation",
+    "hidden_act": "silu",
+    "attention_probs_dropout_prob": 0.0,
+    "bos_token_id": 1,
+    "eos_token_id": 2,
+    "pad_token_id": 0,
+    "intermediate_size": 3072,
+    "initializer_range": 0.02,
+    "layer_norm_eps": 1e-05,
+    "hidden_size": 1024,
+    "attention_head_size": 128,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 18,
+    "num_key_value_heads": 2,
+    "rope_theta": 500000.0,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "vocab_size": 103424,
+    "segment_vocab_size": 0,
+    "skip_init": true,
+    "rope_rank": "adjacent",
+    "max_position_embeddings": 131072,
+    "convert_logits_dtype": "float32",
+    "generation_config": {
+        "tokenizer_config": {
+            "skip_special_tokens": true,
+            "add_special_tokens": false
+        },
+        "eos_token_id": [
+            2
+        ],
+        "max_length": 131072
+    }
+}