Gemma 4 E2B

config.json

@@ -3,6 +3,9 @@
     "Gemma4ForConditionalGeneration"
   ],
   "audio_config": {
+    "_name_or_path": "",
+    "architectures": null,
+    "chunk_size_feed_forward": 0,
     "conf_attention_chunk_size": 12,
     "conf_attention_context_left": 13,
     "conf_attention_context_right": 0,
@@ -19,9 +22,22 @@
     "embedding_norm_eps": 1e-06,
     "gradient_clipping": 10000000000.0,
     "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
     "input_feat_size": 128,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
     "model_type": "gemma4_audio",
+    "output_attentions": false,
+    "output_hidden_states": false,
     "output_proj_dims": 1536,
+    "problem_type": null,
+    "return_dict": true,
     "rms_norm_eps": 1e-06,
     "sscp_conv_channel_size": [
       128,
@@ -56,10 +72,12 @@
     "streaming": false,
     "streaming_chunk_size": 51,
     "streaming_stride": 48,
+    "use_clipped_linears": true,
     "vocab_offset": 262272,
     "vocab_size": 128
   },
-  "audio_soft_tokens_per_image": 188,
+  "audio_ms_per_token": 40,
+  "audio_soft_tokens_per_image": 750,
   "audio_token_id": 258881,
   "boa_token_id": 256000,
   "boi_token_id": 255999,
@@ -78,11 +96,11 @@
     "attn_logit_softcapping": null,
     "bos_token_id": 2,
     "dtype": "bfloat16",
+    "enable_moe_block": false,
     "eos_token_id": 1,
     "expert_intermediate_size": null,
     "final_logit_softcapping": 30.0,
     "global_head_dim": 512,
-    "global_partial_rotary_factor": 0.25,
     "head_dim": 256,
     "hidden_activation": "gelu_pytorch_tanh",
     "hidden_size": 1536,
@@ -135,17 +153,15 @@
     "num_key_value_heads": 1,
     "num_kv_shared_layers": 20,
     "pad_token_id": 0,
-    "partial_rotary_factor": 1.0,
     "query_pre_attn_scalar": 256,
     "rms_norm_eps": 1e-06,
     "rope_parameters": {
       "full_attention": {
-        "partial_rotary_factor": 1.0,
+        "partial_rotary_factor": 0.25,
         "rope_theta": 1000000.0,
         "rope_type": "proportional"
       },
       "sliding_attention": {
-        "partial_rotary_factor": 1.0,
         "rope_theta": 10000.0,
         "rope_type": "default"
       }
@@ -159,7 +175,6 @@
     "use_bidirectional_attention": null,
     "use_cache": true,
     "use_double_wide_mlp": true,
-    "use_second_mlp_block": false,
     "vocab_size": 262144,
     "vocab_size_per_layer_input": 262144
   },

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c17cc59dd92ed94c5d01c2980107d67c06439d21924924d56987dc5ffaaa3f15
-size 10246559478
+oid sha256:6e908419f2891b4cddaff87506a21aa22fce730e9dac81ef1ae59d85d412d8f6
+size 10246627534

processor_config.json

@@ -1,5 +1,29 @@
 {
+  "audio_ms_per_token": 40,
+  "audio_seq_length": 750,
+  "feature_extractor": {
+    "dither": 0.0,
+    "feature_extractor_type": "Gemma4AudioFeatureExtractor",
+    "feature_size": 128,
+    "fft_length": 1024,
+    "fft_overdrive": true,
+    "frame_length": 320,
+    "hop_length": 160,
+    "input_scale_factor": 1.0,
+    "max_frequency": 8000.0,
+    "mel_floor": 0.001,
+    "min_frequency": 0.0,
+    "padding_side": "right",
+    "padding_value": 0.0,
+    "per_bin_mean": null,
+    "per_bin_stddev": null,
+    "preemphasis": 0.0,
+    "preemphasis_htk_flavor": true,
+    "return_attention_mask": true,
+    "sampling_rate": 16000
+  },
   "image_processor": {
+    "data_format": "channels_first",
     "do_convert_rgb": true,
     "do_normalize": false,
     "do_rescale": true,
@@ -9,7 +33,7 @@
       0.5,
       0.5
     ],
-    "image_processor_type": "Gemma4ImageProcessor",
+    "image_processor_type": "Gemma4ImageProcessorFast",
     "image_seq_length": 280,
     "image_std": [
       0.5,

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e9f1033df7d7df8240d1c80089baee77df1980b4131a8be7beb334ec8a48fa2
-size 32169763
+oid sha256:b6f9cb1153f49bc7e0c148d0c1766017aed9512f45eb33afeeb71565d29c938b
+size 32169319