config.json

{
  "_attn_implementation_autoset": true,
  "_name_or_path": "/tmp/tmptlh0zdli",
  "architectures": [
    "CLIPModel"
  ],
  "initializer_factor": 1.0,
  "logit_scale_init_value": 2.6592,
  "model_type": "clip",
  "neuron": {
    "auto_cast": "matmul",
    "auto_cast_type": "bf16",
    "compiler_type": "neuronx-cc",
    "compiler_version": "2.16.372.0+4a9b2326",
    "disable_fallback": false,
    "disable_fast_relayout": false,
    "dynamic_batch_size": false,
    "inline_weights_to_neff": true,
    "input_names": [
      "input_ids",
      "pixel_values",
      "attention_mask"
    ],
    "model_type": "clip",
    "optlevel": "2",
    "output_attentions": false,
    "output_hidden_states": false,
    "output_names": [
      "logits_per_image",
      "logits_per_text",
      "text_embeds",
      "image_embeds",
      "text_model_output",
      "vision_model_output"
    ],
    "static_height": 224,
    "static_image_batch_size": 1,
    "static_num_channels": 3,
    "static_sequence_length": 77,
    "static_text_batch_size": 2,
    "static_width": 224,
    "tensor_parallel_size": 1
  },
  "projection_dim": 512,
  "task": "feature-extraction",
  "text_config": {
    "bos_token_id": 0,
    "dropout": 0.0,
    "eos_token_id": 2,
    "model_type": "clip_text_model",
    "torch_dtype": "float32"
  },
  "torch_dtype": "float32",
  "torchscript": true,
  "transformers_version": "4.49.0",
  "vision_config": {
    "dropout": 0.0,
    "model_type": "clip_vision_model",
    "torch_dtype": "float32"
  }
}