Upload nanoVLM using push_to_hub

config.json

@@ -2,34 +2,34 @@
     "vit_hidden_dim": 768,
     "vit_inter_dim": 3072,
     "vit_patch_size": 16,
-    "vit_img_size": 512,
+    "vit_img_size": 224,
     "vit_n_heads": 12,
     "vit_dropout": 0.0,
     "vit_n_blocks": 12,
     "vit_ln_eps": 1e-06,
     "vit_cls_flag": false,
-    "vit_model_type": "google/siglip2-base-patch16-512",
+    "vit_model_type": "google/siglip2-base-patch16-224",
     "lm_hidden_dim": 960,
     "lm_inter_dim": 2560,
     "lm_rms_eps": 1e-05,
     "lm_re_base": 100000,
     "lm_max_position_embeddings": 8192,
     "lm_base_vocab_size": 49152,
-    "extra_token_amount": 66,
-    "lm_vocab_size": 49218,
+    "extra_token_amount": 102,
+    "lm_vocab_size": 49254,
     "lm_n_heads": 15,
     "lm_n_kv_heads": 5,
     "lm_dropout": 0.0,
     "lm_n_blocks": 32,
     "lm_attn_scaling": 1.0,
-    "lm_max_length": 4096,
+    "lm_max_length": 2048,
     "lm_use_tokens": false,
     "lm_tie_weights": true,
     "lm_model_type": "HuggingFaceTB/SmolLM2-360M-Instruct",
     "lm_tokenizer": "HuggingFaceTB/SmolLM2-360M-Instruct",
     "lm_chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
-    "mp_pixel_shuffle_factor": 4,
-    "mp_image_token_length": 64,
+    "mp_pixel_shuffle_factor": 2,
+    "mp_image_token_length": 49,
     "max_img_size": 2048,
     "resize_to_max_side_len": true,
     "vlm_extra_tokens": {
@@ -43,6 +43,8 @@
         "r1c6": "<row_1_col_6>",
         "r1c7": "<row_1_col_7>",
         "r1c8": "<row_1_col_8>",
+        "r1c9": "<row_1_col_9>",
+        "r1c10": "<row_1_col_10>",
         "r2c1": "<row_2_col_1>",
         "r2c2": "<row_2_col_2>",
         "r2c3": "<row_2_col_3>",
@@ -51,6 +53,8 @@
         "r2c6": "<row_2_col_6>",
         "r2c7": "<row_2_col_7>",
         "r2c8": "<row_2_col_8>",
+        "r2c9": "<row_2_col_9>",
+        "r2c10": "<row_2_col_10>",
         "r3c1": "<row_3_col_1>",
         "r3c2": "<row_3_col_2>",
         "r3c3": "<row_3_col_3>",
@@ -59,6 +63,8 @@
         "r3c6": "<row_3_col_6>",
         "r3c7": "<row_3_col_7>",
         "r3c8": "<row_3_col_8>",
+        "r3c9": "<row_3_col_9>",
+        "r3c10": "<row_3_col_10>",
         "r4c1": "<row_4_col_1>",
         "r4c2": "<row_4_col_2>",
         "r4c3": "<row_4_col_3>",
@@ -67,6 +73,8 @@
         "r4c6": "<row_4_col_6>",
         "r4c7": "<row_4_col_7>",
         "r4c8": "<row_4_col_8>",
+        "r4c9": "<row_4_col_9>",
+        "r4c10": "<row_4_col_10>",
         "r5c1": "<row_5_col_1>",
         "r5c2": "<row_5_col_2>",
         "r5c3": "<row_5_col_3>",
@@ -75,6 +83,8 @@
         "r5c6": "<row_5_col_6>",
         "r5c7": "<row_5_col_7>",
         "r5c8": "<row_5_col_8>",
+        "r5c9": "<row_5_col_9>",
+        "r5c10": "<row_5_col_10>",
         "r6c1": "<row_6_col_1>",
         "r6c2": "<row_6_col_2>",
         "r6c3": "<row_6_col_3>",
@@ -83,6 +93,8 @@
         "r6c6": "<row_6_col_6>",
         "r6c7": "<row_6_col_7>",
         "r6c8": "<row_6_col_8>",
+        "r6c9": "<row_6_col_9>",
+        "r6c10": "<row_6_col_10>",
         "r7c1": "<row_7_col_1>",
         "r7c2": "<row_7_col_2>",
         "r7c3": "<row_7_col_3>",
@@ -91,6 +103,8 @@
         "r7c6": "<row_7_col_6>",
         "r7c7": "<row_7_col_7>",
         "r7c8": "<row_7_col_8>",
+        "r7c9": "<row_7_col_9>",
+        "r7c10": "<row_7_col_10>",
         "r8c1": "<row_8_col_1>",
         "r8c2": "<row_8_col_2>",
         "r8c3": "<row_8_col_3>",
@@ -98,9 +112,31 @@
         "r8c5": "<row_8_col_5>",
         "r8c6": "<row_8_col_6>",
         "r8c7": "<row_8_col_7>",
-        "r8c8": "<row_8_col_8>"
+        "r8c8": "<row_8_col_8>",
+        "r8c9": "<row_8_col_9>",
+        "r8c10": "<row_8_col_10>",
+        "r9c1": "<row_9_col_1>",
+        "r9c2": "<row_9_col_2>",
+        "r9c3": "<row_9_col_3>",
+        "r9c4": "<row_9_col_4>",
+        "r9c5": "<row_9_col_5>",
+        "r9c6": "<row_9_col_6>",
+        "r9c7": "<row_9_col_7>",
+        "r9c8": "<row_9_col_8>",
+        "r9c9": "<row_9_col_9>",
+        "r9c10": "<row_9_col_10>",
+        "r10c1": "<row_10_col_1>",
+        "r10c2": "<row_10_col_2>",
+        "r10c3": "<row_10_col_3>",
+        "r10c4": "<row_10_col_4>",
+        "r10c5": "<row_10_col_5>",
+        "r10c6": "<row_10_col_6>",
+        "r10c7": "<row_10_col_7>",
+        "r10c8": "<row_10_col_8>",
+        "r10c9": "<row_10_col_9>",
+        "r10c10": "<row_10_col_10>"
     },
     "vlm_load_backbone_weights": true,
-    "vlm_checkpoint_path": "checkpoints",
+    "vlm_checkpoint_path": "./checkpoints",
     "hf_repo_name": "nanoVLM"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c7908b2ebefa2e7d5911b2241c4354b82b1fda9978e35a7b2bfe96bec1592f6f
-size 1840504504
+oid sha256:41da8139fed05f1eccf1f7538a6bbef21aa0045f3285c3047489f272ac9b6a92
+size 1802709648