Upload folder using huggingface_hub

README.md

@@ -13,7 +13,7 @@ library_name: transformers
 ---
 
 # EZCon/Qwen2.5-VL-7B-Instruct-4bit-mlx
-This model was converted to MLX format from [`unsloth/Qwen2.5-VL-7B-Instruct`]() using mlx-vlm version **0.3.2**.
+This model was converted to MLX format from [`unsloth/Qwen2.5-VL-7B-Instruct`]() using mlx-vlm version **0.3.3**.
 Refer to the [original model card](https://huggingface.co/unsloth/Qwen2.5-VL-7B-Instruct) for more details on the model.
 ## Use with mlx
 

config.json

@@ -12,6 +12,7 @@
     "decoder_start_token_id": null,
     "diversity_penalty": 0.0,
     "do_sample": false,
+    "dtype": "bfloat16",
     "early_stopping": false,
     "encoder_no_repeat_ngram_size": 0,
     "eos_token_id": 151645,
@@ -36,7 +37,7 @@
     },
     "length_penalty": 1.0,
     "max_length": 20,
-    "max_position_embeddings": 128000,
+    "max_position_embeddings": 32768,
     "max_window_layers": 28,
     "min_length": 0,
     "model_type": "qwen2_5_vl",
@@ -56,11 +57,13 @@
     "pruned_heads": {},
     "quantization": {
         "group_size": 64,
-        "bits": 4
+        "bits": 4,
+        "mode": "affine"
     },
     "quantization_config": {
         "group_size": 64,
-        "bits": 4
+        "bits": 4,
+        "mode": "affine"
     },
     "remove_invalid_values": false,
     "repetition_penalty": 1.0,
@@ -143,7 +146,7 @@
         "return_dict": true,
         "output_hidden_states": false,
         "torchscript": false,
-        "torch_dtype": "bfloat16",
+        "dtype": "bfloat16",
         "pruned_heads": {},
         "tie_word_embeddings": false,
         "chunk_size_feed_forward": 0,
@@ -168,8 +171,8 @@
         "problem_type": null,
         "tokenizer_class": null,
         "prefix": null,
-        "bos_token_id": null,
-        "pad_token_id": 151654,
+        "bos_token_id": 151643,
+        "pad_token_id": null,
         "eos_token_id": 151645,
         "sep_token_id": null,
         "decoder_start_token_id": null,
@@ -178,8 +181,6 @@
         "do_sample": false,
         "early_stopping": false,
         "num_beams": 1,
-        "num_beam_groups": 1,
-        "diversity_penalty": 0.0,
         "temperature": 1.0,
         "top_k": 50,
         "top_p": 1.0,
@@ -198,27 +199,24 @@
         "exponential_decay_length_penalty": null,
         "suppress_tokens": null,
         "begin_suppress_tokens": null,
+        "num_beam_groups": 1,
+        "diversity_penalty": 0.0,
         "_name_or_path": "",
         "model_type": "qwen2_5_vl_text",
-        "unsloth_fixed": true,
         "vision_end_token_id": 151653,
         "vision_start_token_id": 151652,
         "vision_token_id": 151654,
-        "tf_legacy_loss": false,
-        "use_bfloat16": false,
         "output_attentions": false
     },
-    "tf_legacy_loss": false,
     "tie_encoder_decoder": false,
     "tie_word_embeddings": false,
     "tokenizer_class": null,
     "top_k": 50,
     "top_p": 1.0,
     "torchscript": false,
-    "transformers_version": "4.56.0.dev0",
+    "transformers_version": "4.57.0.dev0",
     "typical_p": 1.0,
     "unsloth_fixed": true,
-    "use_bfloat16": false,
     "use_cache": true,
     "use_sliding_window": false,
     "video_token_id": 151656,
@@ -226,7 +224,7 @@
         "return_dict": true,
         "output_hidden_states": false,
         "torchscript": false,
-        "torch_dtype": "bfloat16",
+        "dtype": "bfloat16",
         "pruned_heads": {},
         "tie_word_embeddings": true,
         "chunk_size_feed_forward": 0,
@@ -259,8 +257,6 @@
         "do_sample": false,
         "early_stopping": false,
         "num_beams": 1,
-        "num_beam_groups": 1,
-        "diversity_penalty": 0.0,
         "temperature": 1.0,
         "top_k": 50,
         "top_p": 1.0,
@@ -279,12 +275,12 @@
         "exponential_decay_length_penalty": null,
         "suppress_tokens": null,
         "begin_suppress_tokens": null,
+        "num_beam_groups": 1,
+        "diversity_penalty": 0.0,
         "_name_or_path": "",
         "in_chans": 3,
         "model_type": "qwen2_5_vl",
         "spatial_patch_size": 14,
-        "tf_legacy_loss": false,
-        "use_bfloat16": false,
         "depth": 32,
         "hidden_size": 1280,
         "hidden_act": "silu",

generation_config.json

@@ -5,9 +5,9 @@
     151645,
     151643
   ],
-  "max_length": 128000,
+  "max_length": 32768,
   "pad_token_id": 151654,
   "repetition_penalty": 1.05,
   "temperature": 1e-06,
-  "transformers_version": "4.51.3"
+  "transformers_version": "4.56.1"
 }

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4cf0528116e672daa72b6623dbfd2e006438de499a1821e78aefa22f8f29cc11
+oid sha256:ccc45d5108d1f396b0572e9adac725c5bbc37aacf383a5b97316846e168e6aec
 size 5330941330

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c04dc53c361cb18cf2975990700dee5e92b6414b0930b037c6d56c8107658b3b
+oid sha256:c11d35d4fb4df53bd39c0cc3baebdfd312168ee9a24873cec45aa4847e101255
 size 306561371

model.safetensors.index.json

@@ -1,5 +1,6 @@
 {
   "metadata": {
+    "total_parameters": 8292166656,
     "total_size": 16584333312
   },
   "weight_map": {

preprocessor_config.json

@@ -7,6 +7,7 @@
   "do_center_crop": null,
   "do_convert_rgb": true,
   "do_normalize": true,
+  "do_pad": null,
   "do_rescale": true,
   "do_resize": true,
   "image_mean": [
@@ -24,6 +25,7 @@
   "max_pixels": 12845056,
   "merge_size": 2,
   "min_pixels": 3136,
+  "pad_size": null,
   "patch_size": 14,
   "processor_class": "Qwen2_5_VLProcessor",
   "resample": 3,

tokenizer_config.json

@@ -199,7 +199,7 @@
   "eos_token": "<|im_end|>",
   "errors": "replace",
   "extra_special_tokens": {},
-  "model_max_length": 128000,
+  "model_max_length": 32768,
   "pad_token": "<|vision_pad|>",
   "padding_side": "left",
   "processor_class": "Qwen2_5_VLProcessor",

video_preprocessor_config.json

@@ -28,10 +28,12 @@
   "min_frames": 4,
   "min_pixels": 3136,
   "num_frames": null,
+  "pad_size": null,
   "patch_size": 14,
   "processor_class": "Qwen2_5_VLProcessor",
   "resample": 3,
   "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
   "size": {
     "longest_edge": 12845056,
     "shortest_edge": 3136