Upload folder using huggingface_hub

config.json

@@ -10,7 +10,6 @@
     "conf_conv_kernel_size": 5,
     "conf_num_attention_heads": 8,
     "conf_num_hidden_layers": 12,
-    "conf_positional_bias_size": 256,
     "conf_reduction_factor": 4,
     "conf_residual_weight": 0.5,
     "gradient_clipping": 10000000000.0,
@@ -22,7 +21,7 @@
       128,
       32
     ],
-    "sscp_conv_eps": 0.001,
+    "sscp_conv_group_norm_eps": 0.001,
     "sscp_conv_kernel_size": [
       [
         3,
@@ -51,14 +50,11 @@
   "audio_token_id": 262273,
   "boa_token_id": 256000,
   "boi_token_id": 255999,
-  "bos_token_id": 2,
   "eoa_token_id": 262272,
   "eoi_token_id": 262144,
-  "eos_token_id": 1,
   "image_token_id": 262145,
   "initializer_range": 0.02,
   "model_type": "gemma3n",
-  "pad_token_id": 0,
   "text_config": {
     "activation_sparsity_pattern": [
       0.95,
@@ -100,7 +96,6 @@
     "altup_active_idx": 0,
     "altup_coef_clip": 120.0,
     "altup_correct_scale": true,
-    "altup_lr_multiplier": 1.0,
     "altup_num_inputs": 4,
     "attention_bias": false,
     "attention_dropout": 0.0,
@@ -110,7 +105,43 @@
     "hidden_size": 2048,
     "hidden_size_per_layer_input": 256,
     "initializer_range": 0.02,
-    "intermediate_size": 16384,
+    "intermediate_size": [
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384,
+      16384
+    ],
     "laurel_rank": 64,
     "layer_types": [
       "sliding_attention",
@@ -155,7 +186,6 @@
     "num_hidden_layers": 35,
     "num_key_value_heads": 2,
     "num_kv_shared_layers": 15,
-    "query_pre_attn_scalar": 256,
     "rms_norm_eps": 1e-06,
     "rope_local_base_freq": 10000.0,
     "rope_scaling": null,
@@ -167,17 +197,17 @@
     "vocab_size_per_layer_input": 262144
   },
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.53.0.dev0",
-  "unsloth_fixed": true,
+  "transformers_version": "4.54.0.dev0",
   "vision_config": {
     "architecture": "mobilenetv5_300m_enc",
-    "do_pooling": true,
+    "do_pooling": false,
     "hidden_size": 2048,
     "initializer_range": 0.02,
     "label_names": [
       "LABEL_0",
       "LABEL_1"
     ],
+    "model_args": null,
     "model_type": "gemma3n_vision",
     "num_classes": 2,
     "rms_norm_eps": 1e-06,

generation_config.json

@@ -6,5 +6,5 @@
   "pad_token_id": 0,
   "top_k": 64,
   "top_p": 0.95,
-  "transformers_version": "4.53.0.dev0"
+  "transformers_version": "4.54.0.dev0"
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:681e3230979df3dcf723631523de45d7bdd69be8af6e1c348c2ceeffe17595f9
+oid sha256:6dc80317bce6640e4ca4e674b3889eedf26e723f1fbeb956289ab1ef01c21520
 size 3077103576

preprocessor_config.json

@@ -3,12 +3,14 @@
   "data_format": "channels_first",
   "default_to_square": false,
   "device": null,
+  "disable_grouping": null,
   "dither": 0.0,
   "do_center_crop": null,
   "do_convert_rgb": null,
   "do_normalize": false,
   "do_rescale": true,
   "do_resize": true,
+  "feature_extractor_type": "Gemma3nAudioFeatureExtractor",
   "feature_size": 128,
   "fft_length": 1024,
   "fft_overdrive": true,
@@ -31,7 +33,7 @@
   "max_frequency": 7600.0,
   "mel_floor": 1e-05,
   "min_frequency": 125.0,
-  "padding_side": "left",
+  "padding_side": "right",
   "padding_value": 0.0,
   "per_bin_mean": null,
   "per_bin_stddev": null,
@@ -40,7 +42,7 @@
   "processor_class": "Gemma3nProcessor",
   "resample": 2,
   "rescale_factor": 0.00392156862745098,
-  "return_attention_mask": false,
+  "return_attention_mask": true,
   "return_tensors": null,
   "sampling_rate": 16000,
   "size": {

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b6c35ee648c07754b44cd9e371c75d4caa05c4504910b7ad29b1847ee9d8ba5d
-size 33442553
+oid sha256:c4c19736bf24d1c6805cf49340e31bd02c70fb7857a2cb31065c90c2b5719c4e
+size 33442559

tokenizer_config.json

@@ -2160,7 +2160,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": true
+      "special": false
     },
     "256000": {
       "content": "<start_of_audio>",
@@ -2168,7 +2168,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": true
+      "special": false
     },
     "256001": {
       "content": "<unused99>",
@@ -51320,7 +51320,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": true
+      "special": false
     },
     "262145": {
       "content": "<image_soft_token>",
@@ -51328,7 +51328,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": true
+      "special": false
     },
     "262146": {
       "content": "<unused_image0>",
@@ -52344,7 +52344,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": true
+      "special": false
     },
     "262273": {
       "content": "<audio_soft_token>",
@@ -52352,7 +52352,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": true
+      "special": false
     },
     "262274": {
       "content": "<unused_audio0>",
@@ -53380,13 +53380,12 @@
     "image_token": "<image_soft_token>"
   },
   "image_token": "<image_soft_token>",
-  "model_max_length": 32768,
+  "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<pad>",
-  "padding_side": "left",
   "processor_class": "Gemma3nProcessor",
   "sp_model_kwargs": null,
   "spaces_between_special_tokens": false,
   "tokenizer_class": "GemmaTokenizer",
   "unk_token": "<unk>",
   "use_default_system_prompt": false
-}
+}