Add SmolVLM2-500M sidecar pipeline (DepthBridge + ObjectAnchorProjector)

README.md

@@ -0,0 +1,41 @@
+---
+license: apache-2.0
+base_model: HuggingFaceTB/SmolVLM2-500M-Video-Instruct
+tags:
+  - smolvlm
+  - depth-estimation
+  - object-detection
+  - multimodal
+  - sidecar-pipeline
+---
+
+# SmolVLM2-500M-Video-Instruct + Sidecar Pipeline
+
+This model extends [SmolVLM2-500M-Video-Instruct](HuggingFaceTB/SmolVLM2-500M-Video-Instruct) with two
+lightweight sidecar modules for grounded spatial reasoning:
+
+| Module | Params | Purpose |
+|---|---|---|
+| **DepthBridge** | ~760 K | Fuses Depth-Anything-V2-Metric depth maps into SigLIP patch tokens via a gated residual |
+| **ObjectAnchorProjector** | ~1.3 K | Projects YOLOv8-World CLIP embeddings into LM anchor tokens |
+
+## Sidecar config flags
+
+```python
+config.depth_integration  = True   # enables DepthBridge
+config.object_integration = True   # enables ObjectAnchorProjector (train first)
+```
+
+## Loading
+
+```python
+from transformers import AutoProcessor, AutoModelForImageTextToText
+
+model     = AutoModelForImageTextToText.from_pretrained("anuragpradhan/SmolVLM2-500M-DepthAwareVLM")
+processor = AutoProcessor.from_pretrained("anuragpradhan/SmolVLM2-500M-DepthAwareVLM")
+```
+
+## Fine-tuning note
+
+DepthBridge gate α is initialised at 0.0 (depth is inactive until trained).
+Run `model.freeze_base_models()` to train only the sidecar modules.

chat_template.jinja

@@ -0,0 +1,2 @@
+<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
+{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}

config.json

@@ -0,0 +1,168 @@
+{
+  "architectures": [
+    "SmolVLMForConditionalGeneration"
+  ],
+  "depth_gate_init": 0.0,
+  "depth_hidden_dim": 256,
+  "depth_integration": true,
+  "dtype": "float32",
+  "image_token_id": 49190,
+  "max_objects": 20,
+  "model_type": "smolvlm",
+  "object_feature_dim": 512,
+  "object_integration": true,
+  "pad_token_id": 128002,
+  "scale_factor": 4,
+  "text_config": {
+    "_flash_attn_2_enabled": true,
+    "_name_or_path": "None",
+    "architectures": [
+      "VLlama3ForCausalLM"
+    ],
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 1,
+    "dtype": "float32",
+    "eos_token_id": 2,
+    "head_dim": 64,
+    "hidden_act": "silu",
+    "hidden_size": 960,
+    "initializer_range": 0.02,
+    "intermediate_size": 2560,
+    "is_llama_config": true,
+    "max_position_embeddings": 8192,
+    "mlp_bias": false,
+    "model_type": "llama",
+    "neftune_noise_alpha": 0.0,
+    "num_attention_heads": 15,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 5,
+    "pad_token_id": 2,
+    "perceiver_config": {
+      "_attn_implementation_autoset": false,
+      "_name_or_path": "",
+      "add_cross_attention": false,
+      "architectures": null,
+      "attention_dropout": 0.0,
+      "bad_words_ids": null,
+      "begin_suppress_tokens": null,
+      "bos_token_id": null,
+      "chunk_size_feed_forward": 0,
+      "cross_attention_hidden_size": null,
+      "decoder_start_token_id": null,
+      "diversity_penalty": 0.0,
+      "do_sample": false,
+      "early_stopping": false,
+      "encoder_no_repeat_ngram_size": 0,
+      "eos_token_id": null,
+      "exponential_decay_length_penalty": null,
+      "finetuning_task": null,
+      "forced_bos_token_id": null,
+      "forced_eos_token_id": null,
+      "hidden_act": "silu",
+      "id2label": {
+        "0": "LABEL_0",
+        "1": "LABEL_1"
+      },
+      "is_decoder": false,
+      "is_encoder_decoder": false,
+      "label2id": {
+        "LABEL_0": 0,
+        "LABEL_1": 1
+      },
+      "length_penalty": 1.0,
+      "max_length": 20,
+      "min_length": 0,
+      "model_type": "vllama3",
+      "no_repeat_ngram_size": 0,
+      "num_beam_groups": 1,
+      "num_beams": 1,
+      "num_key_value_heads": 1,
+      "num_return_sequences": 1,
+      "output_attentions": false,
+      "output_hidden_states": false,
+      "output_scores": false,
+      "pad_token_id": null,
+      "prefix": null,
+      "problem_type": null,
+      "pruned_heads": {},
+      "qk_layer_norms_perceiver": false,
+      "remove_invalid_values": false,
+      "repetition_penalty": 1.0,
+      "resampler_depth": 6,
+      "resampler_head_dim": 96,
+      "resampler_n_heads": 16,
+      "resampler_n_latents": 64,
+      "return_dict": true,
+      "return_dict_in_generate": false,
+      "sep_token_id": null,
+      "suppress_tokens": null,
+      "task_specific_params": null,
+      "temperature": 1.0,
+      "tf_legacy_loss": false,
+      "tie_encoder_decoder": false,
+      "tie_word_embeddings": true,
+      "tokenizer_class": null,
+      "top_k": 50,
+      "top_p": 1.0,
+      "torch_dtype": null,
+      "torchscript": false,
+      "transformers_version": "4.46.0",
+      "typical_p": 1.0,
+      "use_bfloat16": false
+    },
+    "pixel_shuffle_factor": 4,
+    "pretraining_tp": 1,
+    "qk_layer_norms": false,
+    "rms_norm_eps": 1e-05,
+    "rope_interleaved": false,
+    "rope_parameters": {
+      "rope_theta": 100000,
+      "rope_type": "default"
+    },
+    "tie_word_embeddings": false,
+    "transformers.js_config": {
+      "kv_cache_dtype": {
+        "fp16": "float16",
+        "q4f16": "float16"
+      }
+    },
+    "use_cache": true,
+    "use_resampler": false,
+    "vocab_size": 49280
+  },
+  "tie_word_embeddings": false,
+  "transformers.js_config": {
+    "kv_cache_dtype": {
+      "fp16": "float16",
+      "q4f16": "float16"
+    }
+  },
+  "transformers_version": "5.8.0.dev0",
+  "use_cache": false,
+  "use_reentrant_checkpointing": false,
+  "vision_config": {
+    "attention_dropout": 0.0,
+    "dtype": "float32",
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 768,
+    "image_size": 512,
+    "initializer_range": 0.02,
+    "intermediate_size": 3072,
+    "layer_norm_eps": 1e-06,
+    "max_image_size": {
+      "longest_edge": 512
+    },
+    "model_type": "smolvlm_vision",
+    "num_attention_heads": 12,
+    "num_channels": 3,
+    "num_hidden_layers": 12,
+    "patch_size": 16,
+    "size": {
+      "longest_edge": 2048
+    },
+    "tie_word_embeddings": false,
+    "use_base_siglip": false
+  },
+  "vocab_size": 49280
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 49279,
+  "pad_token_id": 2,
+  "transformers_version": "5.8.0.dev0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fad09e4f72fe9b68a9d96c49ce3076346137f8a6e996973cbe73a1c3cab241bf
+size 2033036156

processor_config.json

@@ -0,0 +1,76 @@
+{
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_image_splitting": true,
+    "do_normalize": true,
+    "do_pad": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "SmolVLMImageProcessor",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_image_size": {
+      "longest_edge": 512
+    },
+    "resample": 1,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "longest_edge": 2048
+    },
+    "video_sampling": {
+      "fps": 1,
+      "max_frames": 64,
+      "video_size": {
+        "longest_edge": 512
+      }
+    }
+  },
+  "image_seq_len": 64,
+  "processor_class": "SmolVLMProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_image_splitting": true,
+    "do_normalize": true,
+    "do_pad": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": false,
+    "fps": 1,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_image_size": {
+      "longest_edge": 512
+    },
+    "num_frames": 64,
+    "resample": 1,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "size": {
+      "longest_edge": 2048
+    },
+    "video_processor_type": "SmolVLMVideoProcessor",
+    "video_sampling": {
+      "fps": 1,
+      "max_frames": 64,
+      "video_size": {
+        "longest_edge": 2048
+      }
+    }
+  }
+}

tokenizer_config.json

@@ -0,0 +1,28 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": "<|im_start|>",
+  "clean_up_tokenization_spaces": false,
+  "end_of_utterance_token": "<end_of_utterance>",
+  "eos_token": "<end_of_utterance>",
+  "errors": "replace",
+  "fake_image_token": "<fake_token_around_image>",
+  "global_image_token": "<global-img>",
+  "image_token": "<image>",
+  "is_local": false,
+  "legacy": false,
+  "local_files_only": false,
+  "model_max_length": 8192,
+  "model_specific_special_tokens": {
+    "end_of_utterance_token": "<end_of_utterance>",
+    "fake_image_token": "<fake_token_around_image>",
+    "global_image_token": "<global-img>",
+    "image_token": "<image>"
+  },
+  "pad_token": "<|im_end|>",
+  "processor_class": "SmolVLMProcessor",
+  "tokenizer_class": "GPT2Tokenizer",
+  "truncation_side": "left",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}