Upload folder using huggingface_hub

README.md

@@ -1,11 +1,10 @@
 ---
+license: apache-2.0
+language:
+- en
 base_model:
 - openai/clip-vit-large-patch14-336
 - Qwen/Qwen2-7B
-language:
-- en
-library_name: transformers
-license: apache-2.0
 pipeline_tag: image-text-to-text
 tags:
 - multimodal
@@ -13,10 +12,11 @@ tags:
 - molmo
 - pixmo
 - mlx
+library_name: transformers
 ---
 
 # mlx-community/Molmo-7B-D-0924-4bit
-This model was converted to MLX format from [`allenai/Molmo-7B-D-0924`]() using mlx-vlm version **0.1.0**.
+This model was converted to MLX format from [`allenai/Molmo-7B-D-0924`]() using mlx-vlm version **0.1.6**.
 Refer to the [original model card](https://huggingface.co/allenai/Molmo-7B-D-0924) for more details on the model.
 ## Use with mlx
 

config.json

@@ -1,4 +1,6 @@
 {
+    "_attn_implementation_autoset": false,
+    "add_cross_attention": false,
     "architectures": [
         "MolmoForCausalLM"
     ],
@@ -7,33 +9,145 @@
         "AutoConfig": "config_molmo.MolmoConfig",
         "AutoModelForCausalLM": "modeling_molmo.MolmoForCausalLM"
     },
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
     "clip_qkv": null,
-    "embedding_size": 152064,
-    "hidden_size": 3584,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "embedding_size": 50304,
+    "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_size": 4096,
+    "id2label": {
+        "0": "LABEL_0",
+        "1": "LABEL_1"
+    },
     "initializer_range": 0.02,
-    "intermediate_size": 37888,
-    "layer_norm_eps": 1e-06,
+    "intermediate_size": 11008,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+        "LABEL_0": 0,
+        "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
     "layer_norm_type": "rms",
-    "max_position_embeddings": 4096,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 2048,
+    "min_length": 0,
     "model_type": "molmo",
+    "no_repeat_ngram_size": 0,
     "norm_after": false,
-    "num_attention_heads": 28,
-    "num_hidden_layers": 28,
-    "num_key_value_heads": 4,
-    "qkv_bias": true,
+    "num_attention_heads": 32,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": null,
+    "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": {},
+    "qkv_bias": false,
     "quantization": {
         "group_size": 64,
         "bits": 4
     },
-    "rope_theta": 1000000.0,
-    "tie_word_embeddings": false,
-    "torch_dtype": "float32",
-    "transformers_version": "4.43.3",
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "rope_theta": 10000.0,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "text_config": {
+        "model_type": "molmo",
+        "max_position_embeddings": 4096,
+        "d_model": 3584,
+        "n_heads": 28,
+        "n_kv_heads": 4,
+        "n_layers": 28,
+        "mlp_ratio": 4,
+        "max_sequence_length": 1024,
+        "act_output_multiplier": 0.5,
+        "mlp_hidden_size": 37888,
+        "vocab_size": 152064,
+        "embedding_size": 152064,
+        "additional_vocab_size": 128,
+        "attention_dropout": 0.1,
+        "residual_dropout": 0.1,
+        "embedding_dropout": 0.1,
+        "layer_norm_eps": 1e-05,
+        "initializer_range": 0.02,
+        "pad_token_id": -1,
+        "rope": true,
+        "rope_theta": 1000000.0,
+        "weight_tying": false,
+        "rope_full_precision": true,
+        "rope_impl": "interleave"
+    },
+    "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": "4.47.1",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
     "use_cache": true,
     "use_position_ids": true,
     "vision_config": {
-        "skip_vision_non_divisible": true,
-        "intermediate_size": 37888
+        "model_type": "molmo",
+        "num_channels": 3,
+        "image_default_input_size": [
+            336,
+            336
+        ],
+        "image_patch_size": 14,
+        "image_pos_patch_size": 14,
+        "hidden_size": 18944,
+        "image_emb_dim": 1024,
+        "image_num_heads": 16,
+        "image_num_key_value_heads": 16,
+        "image_num_layers": 23,
+        "image_head_dim": 64,
+        "image_mlp_dim": 4096,
+        "image_mlp_activations": "gelu",
+        "image_dropout_rate": 0.0,
+        "image_num_pos": 577,
+        "image_norm_eps": 1e-05,
+        "attention_dropout": 0.0,
+        "residual_dropout": 0.0,
+        "initializer_range": 0.02,
+        "d_model": 3584,
+        "image_pooling_h": 2,
+        "image_pooling_w": 2,
+        "vit_layers": [
+            -2,
+            -9
+        ],
+        "image_pooling_2d": "attention-meanq",
+        "image_padding_embed": "pad_and_partial_pad",
+        "intermediate_size": 588,
+        "skip_vision_non_divisible": true
     },
     "vocab_size": 152064,
     "weight_tying": false

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c899869c6f30c19c165233fdc36d76421002fcefd72b619c7fd7aaa539399205
+oid sha256:5eff4a266d41eb43d4b4e3a207fc222bfed68a78ca60c54df8c5f285fcfd1294
 size 5298715127

tokenizer_config.json

@@ -3843,6 +3843,7 @@
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|endoftext|>",
   "errors": "replace",
+  "extra_special_tokens": {},
   "model_max_length": 32768,
   "pad_token": "<|endoftext|>",
   "processor_class": "MolmoProcessor",