Upload 10 files

config.json

@@ -1,55 +1,173 @@
 {
-  "_name_or_path": "rkazants/tiny-random-MiniCPM-o-2_6",
+  "_name_or_path": "openbmb/MiniCPM-o-2_6",
   "architectures": [
     "MiniCPMO"
   ],
   "attention_dropout": 0.0,
   "audio_chunk_length": 1.0,
   "audio_config": {
-    "classifier_proj_size": 64,
-    "d_model": 4,
-    "decoder_attention_heads": 2,
-    "decoder_ffn_dim": 32,
-    "decoder_layers": 2,
-    "encoder_attention_heads": 2,
-    "encoder_ffn_dim": 32,
+    "_name_or_path": "openai/whisper-medium",
+    "architectures": [
+      "MiniCPMWhisperEncoder"
+    ],
+    "begin_suppress_tokens": [
+      220,
+      50257
+    ],
+    "bos_token_id": 50257,
+    "d_model": 1024,
+    "decoder_attention_heads": 16,
+    "decoder_ffn_dim": 4096,
+    "decoder_layers": 24,
+    "decoder_start_token_id": 50258,
+    "encoder_attention_heads": 16,
+    "encoder_ffn_dim": 4096,
     "encoder_layers": 2,
-    "intermediate_size": 64,
+    "eos_token_id": 50257,
+    "forced_decoder_ids": [
+      [
+        1,
+        50259
+      ],
+      [
+        2,
+        50359
+      ],
+      [
+        3,
+        50363
+      ]
+    ],
+    "max_length": 448,
     "model_type": "whisper",
-    "num_hidden_layers": 2,
-    "num_mel_bins": 20
+    "num_hidden_layers": 24,
+    "pad_token_id": 50257,
+    "suppress_tokens": [
+      1,
+      2,
+      7,
+      8,
+      9,
+      10,
+      14,
+      25,
+      26,
+      27,
+      28,
+      29,
+      31,
+      58,
+      59,
+      60,
+      61,
+      62,
+      63,
+      90,
+      91,
+      92,
+      93,
+      359,
+      503,
+      522,
+      542,
+      873,
+      893,
+      902,
+      918,
+      922,
+      931,
+      1350,
+      1853,
+      1982,
+      2460,
+      2627,
+      3246,
+      3253,
+      3268,
+      3536,
+      3846,
+      3961,
+      4183,
+      4667,
+      6585,
+      6647,
+      7273,
+      9061,
+      9383,
+      10428,
+      10929,
+      11938,
+      12033,
+      12331,
+      12562,
+      13793,
+      14157,
+      14635,
+      15265,
+      15618,
+      16553,
+      16604,
+      18362,
+      18956,
+      20075,
+      21675,
+      22520,
+      26130,
+      26161,
+      26435,
+      28279,
+      29464,
+      31650,
+      32302,
+      32470,
+      36865,
+      42863,
+      47425,
+      49870,
+      50254,
+      50258,
+      50358,
+      50359,
+      50360,
+      50361,
+      50362
+    ],
+    "torch_dtype": "float32"
   },
   "audio_pool_step": 2,
   "auto_map": {
-    "AutoConfig": "configuration_minicpm.MiniCPMOConfig",
-    "AutoModel": "modeling_minicpmo.MiniCPMO",
-    "AutoModelForCausalLM": "modeling_minicpmo.MiniCPMO"
+    "AutoConfig": "openbmb/MiniCPM-o-2_6--configuration_minicpm.MiniCPMOConfig",
+    "AutoModel": "openbmb/MiniCPM-o-2_6--modeling_minicpmo.MiniCPMO",
+    "AutoModelForCausalLM": "openbmb/MiniCPM-o-2_6--modeling_minicpmo.MiniCPMO"
   },
   "batch_vision_input": true,
-  "drop_vision_last_layer": true,
+  "bos_token_id": 151643,
+  "chunk_input": true,
+  "drop_vision_last_layer": false,
+  "eos_token_id": 151645,
   "hidden_act": "silu",
-  "hidden_size": 128,
-  "image_size": 28,
+  "hidden_size": 168,
+  "image_size": 448,
   "init_audio": true,
   "init_tts": true,
   "init_vision": true,
   "initializer_range": 0.02,
-  "intermediate_size": 128,
+  "intermediate_size": 32,
+  "listen_speak_type": "asr",
   "max_position_embeddings": 32768,
   "max_window_layers": 28,
   "model_type": "minicpmo",
-  "num_attention_heads": 2,
+  "num_attention_heads": 28,
+  "num_heads": 2,
   "num_hidden_layers": 2,
-  "num_key_value_heads": 32,
-  "patch_size": 2,
-  "query_num": 2,
+  "num_key_value_heads": 4,
+  "patch_size": 14,
+  "query_num": 64,
   "rms_norm_eps": 1e-06,
-  "rope_theta": 10000.0,
+  "rope_theta": 1000000.0,
   "slice_config": {
-    "max_slice_nums": 3,
-    "model_type": "minicpmv",
-    "patch_size": 2,
-    "scale_resolution": 256
+    "max_slice_nums": 9,
+    "model_type": "minicpmv"
   },
   "slice_mode": true,
   "sliding_window": null,
@@ -58,26 +176,23 @@
   "torch_dtype": "float32",
   "transformers_version": "4.44.2",
   "tts_config": {
-    "hidden_size": 64,
-    "intermediate_size": 64,
-    "llm_dim": 64,
-    "model_type": "conditional_chattts",
-    "num_attention_heads": 2,
-    "num_hidden_layers": 2,
-    "num_mel_bins": 10
+    "hidden_size": 24,
+    "llm_dim": 32,
+    "model_type": "conditional_chattts"
   },
   "use_cache": true,
   "use_image_id": true,
   "use_sliding_window": false,
-  "vision_batch_size": 2,
+  "version": 2.6,
+  "vision_batch_size": 16,
   "vision_config": {
-    "hidden_size": 64,
-    "image_size": 28,
-    "intermediate_size": 64,
+    "hidden_size": 32,
+    "image_size": 224,
+    "intermediate_size": 32,
     "model_type": "siglip_vision_model",
     "num_attention_heads": 2,
-    "num_hidden_layers": 2,
-    "patch_size": 2
+    "num_hidden_layers": 1,
+    "patch_size": 14
   },
-  "vocab_size": 151936
+  "vocab_size": 151700
 }

generation_config.json

@@ -1,4 +1,6 @@
 {
   "_from_model_config": true,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
   "transformers_version": "4.44.2"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9a72d8aaf4796fa246d22d26940d4601214eb76232423abb960dea88b8fbc607
-size 236425096
+oid sha256:e78bc28359ce3fdf08cfc2dc701d4903c34ebb5943e12decb3d5ce6ad9901f80
+size 410265248

tokenizer_config.json

@@ -506,8 +506,8 @@
   ],
   "auto_map": {
     "AutoTokenizer": [
-      "tokenization_qwen2.Qwen2Tokenizer",
-      "tokenization_minicpmo_fast.MiniCPMOTokenizerFast"
+      "openbmb/MiniCPM-o-2_6--tokenization_minicpmo_fast.MiniCPMOTokenizerFast",
+      null
     ]
   },
   "bos_token": "<|im_start|>",