Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

config.json

@@ -0,0 +1,52 @@
+{
+    "_name_or_path": "nintwentydo/Razorback-12B-v0.2",
+    "architectures": [
+        "LlavaForConditionalGeneration"
+    ],
+    "ignore_index": -100,
+    "image_seq_length": 1,
+    "image_token_index": 10,
+    "model_type": "llava",
+    "projector_hidden_act": "gelu",
+    "text_config": {
+        "hidden_size": 5120,
+        "head_dim": 128,
+        "intermediate_size": 14336,
+        "is_composition": true,
+        "max_position_embeddings": 131072,
+        "model_type": "mistral",
+        "num_hidden_layers": 40,
+        "num_attention_heads": 32,
+        "num_key_value_heads": 8,
+        "rms_norm_eps": 1e-05,
+        "rope_theta": 1000000000.0,
+        "sliding_window": null,
+        "vocab_size": 131072
+    },
+    "torch_dtype": "bfloat16",
+    "transformers_version": "4.47.1",
+    "vision_config": {
+        "head_dim": 64,
+        "num_attention_heads": 16,
+        "hidden_act": "silu",
+        "image_size": 1024,
+        "is_composition": true,
+        "model_type": "pixtral",
+        "patch_size": 16,
+        "rope_theta": 10000.0,
+        "tie_word_embeddings": false
+    },
+    "vision_feature_layer": -1,
+    "vision_feature_select_strategy": "full",
+    "quantization_config": {
+        "quant_method": "exl2",
+        "version": "0.2.7",
+        "bits": 8.0,
+        "head_bits": 8,
+        "calibration": {
+            "rows": 115,
+            "length": 2048,
+            "dataset": "(default)"
+        }
+    }
+}

mergekit_config.yml

@@ -0,0 +1,738 @@
+merge_method: task_arithmetic
+dtype: bfloat16
+base_model: /opt/dlami/nvme/models/unsloth_Pixtral-12B-2409
+models:
+- model: /opt/dlami/nvme/models/TheDrummer_UnslopNemo-12B-v3_llava
+  parameters:
+    normalize: false
+    weight:
+    - filter: language_model.lm_head.weight
+      value: 0.027243
+    - filter: language_model.model.embed_tokens.weight
+      value: 0.14717
+    - filter: language_model.model.layers.39.mlp.gate_proj.weight
+      value: 0.128203
+    - filter: language_model.model.layers.39.mlp.up_proj.weight
+      value: 0.123174
+    - filter: language_model.model.layers.39.mlp.down_proj.weight
+      value: 0.098388
+    - filter: language_model.model.layers.38.mlp.down_proj.weight
+      value: 0.125974
+    - filter: language_model.model.layers.38.mlp.up_proj.weight
+      value: 0.151487
+    - filter: language_model.model.layers.38.mlp.gate_proj.weight
+      value: 0.154605
+    - filter: language_model.model.layers.37.mlp.gate_proj.weight
+      value: 0.160597
+    - filter: language_model.model.layers.37.mlp.up_proj.weight
+      value: 0.160597
+    - filter: language_model.model.layers.37.mlp.down_proj.weight
+      value: 0.154017
+    - filter: language_model.model.layers.36.mlp.up_proj.weight
+      value: 0.166257
+    - filter: language_model.model.layers.36.mlp.gate_proj.weight
+      value: 0.165639
+    - filter: language_model.model.layers.36.mlp.down_proj.weight
+      value: 0.165014
+    - filter: language_model.model.layers.35.mlp.gate_proj.weight
+      value: 0.17496
+    - filter: language_model.model.layers.35.mlp.up_proj.weight
+      value: 0.170804
+    - filter: language_model.model.layers.35.mlp.down_proj.weight
+      value: 0.170191
+    - filter: language_model.model.layers.4.mlp.gate_proj.weight
+      value: 0.186816
+    - filter: language_model.model.layers.6.mlp.gate_proj.weight
+      value: 0.186816
+    - filter: language_model.model.norm.weight
+      value: 0.700371
+    - filter: language_model.model.layers.34.mlp.gate_proj.weight
+      value: 0.175869
+    - filter: language_model.model.layers.7.mlp.gate_proj.weight
+      value: 0.184212
+    - filter: language_model.model.layers.33.mlp.gate_proj.weight
+      value: 0.177382
+    - filter: language_model.model.layers.34.mlp.up_proj.weight
+      value: 0.171061
+    - filter: language_model.model.layers.5.mlp.gate_proj.weight
+      value: 0.187836
+    - filter: language_model.model.layers.8.mlp.gate_proj.weight
+      value: 0.177382
+    - filter: language_model.model.layers.11.mlp.gate_proj.weight
+      value: 0.180445
+    - filter: language_model.model.layers.33.mlp.up_proj.weight
+      value: 0.171948
+    - filter: language_model.model.layers.34.mlp.down_proj.weight
+      value: 0.174078
+    - filter: language_model.model.layers.9.mlp.gate_proj.weight
+      value: 0.176204
+    - filter: language_model.model.layers.10.mlp.gate_proj.weight
+      value: 0.175606
+    - filter: language_model.model.layers.3.mlp.gate_proj.weight
+      value: 0.186236
+    - filter: language_model.model.layers.32.mlp.gate_proj.weight
+      value: 0.184118
+    - filter: language_model.model.layers.12.mlp.gate_proj.weight
+      value: 0.183562
+    - filter: language_model.model.layers.13.mlp.gate_proj.weight
+      value: 0.180853
+    - filter: language_model.model.layers.31.mlp.gate_proj.weight
+      value: 0.189408
+    - filter: language_model.model.layers.32.mlp.up_proj.weight
+      value: 0.172229
+    - filter: language_model.model.layers.33.mlp.down_proj.weight
+      value: 0.176549
+    - filter: language_model.model.layers.14.mlp.gate_proj.weight
+      value: 0.175944
+    - filter: language_model.model.layers.30.mlp.gate_proj.weight
+      value: 0.194776
+    - filter: language_model.model.layers.29.mlp.gate_proj.weight
+      value: 0.194269
+    - filter: language_model.model.layers.31.mlp.up_proj.weight
+      value: 0.174714
+    - filter: language_model.model.layers.32.mlp.down_proj.weight
+      value: 0.179092
+    - filter: language_model.model.layers.15.mlp.gate_proj.weight
+      value: 0.174089
+    - filter: language_model.model.layers.21.mlp.gate_proj.weight
+      value: 0.187243
+    - filter: language_model.model.layers.22.mlp.gate_proj.weight
+      value: 0.189419
+    - filter: language_model.model.layers.27.mlp.gate_proj.weight
+      value: 0.195916
+    - filter: language_model.model.layers.28.mlp.gate_proj.weight
+      value: 0.195916
+    - filter: language_model.model.layers.16.mlp.gate_proj.weight
+      value: 0.173457
+    - filter: language_model.model.layers.2.mlp.gate_proj.weight
+      value: 0.191059
+    - filter: language_model.model.layers.20.mlp.gate_proj.weight
+      value: 0.184494
+    - filter: language_model.model.layers.26.mlp.gate_proj.weight
+      value: 0.19541
+    - filter: language_model.model.layers.30.mlp.up_proj.weight
+      value: 0.175673
+    - filter: language_model.model.layers.18.mlp.gate_proj.weight
+      value: 0.179492
+    - filter: language_model.model.layers.19.mlp.gate_proj.weight
+      value: 0.181707
+    - filter: language_model.model.layers.23.mlp.gate_proj.weight
+      value: 0.192713
+    - filter: language_model.model.layers.24.mlp.gate_proj.weight
+      value: 0.194899
+    - filter: language_model.model.layers.25.mlp.gate_proj.weight
+      value: 0.194899
+    - filter: language_model.model.layers.29.mlp.up_proj.weight
+      value: 0.177272
+    - filter: language_model.model.layers.31.mlp.down_proj.weight
+      value: 0.181707
+    - filter: language_model.model.layers.17.mlp.gate_proj.weight
+      value: 0.176651
+    - filter: language_model.model.layers.28.mlp.up_proj.weight
+      value: 0.176651
+    - filter: language_model.model.layers.26.mlp.up_proj.weight
+      value: 0.179905
+    - filter: language_model.model.layers.27.mlp.up_proj.weight
+      value: 0.179905
+    - filter: language_model.model.layers.30.mlp.down_proj.weight
+      value: 0.182155
+    - filter: language_model.model.layers.1.mlp.gate_proj.weight
+      value: 0.195027
+    - filter: language_model.model.layers.23.mlp.up_proj.weight
+      value: 0.186066
+    - filter: language_model.model.layers.24.mlp.up_proj.weight
+      value: 0.183812
+    - filter: language_model.model.layers.25.mlp.up_proj.weight
+      value: 0.183812
+    - filter: language_model.model.layers.15.mlp.up_proj.weight
+      value: 0.19674
+    - filter: language_model.model.layers.16.mlp.up_proj.weight
+      value: 0.198974
+    - filter: language_model.model.layers.17.mlp.up_proj.weight
+      value: 0.201202
+    - filter: language_model.model.layers.21.mlp.up_proj.weight
+      value: 0.187747
+    - filter: language_model.model.layers.22.mlp.up_proj.weight
+      value: 0.185485
+    - filter: language_model.model.layers.0.mlp.gate_proj.weight
+      value: 0.191707
+    - filter: language_model.model.layers.14.mlp.up_proj.weight
+      value: 0.193966
+    - filter: language_model.model.layers.18.mlp.up_proj.weight
+      value: 0.200708
+    - filter: language_model.model.layers.19.mlp.up_proj.weight
+      value: 0.19622
+    - filter: language_model.model.layers.20.mlp.up_proj.weight
+      value: 0.191707
+    - filter: language_model.model.layers.29.mlp.down_proj.weight
+      value: 0.187172
+    - filter: language_model.model.layers.10.mlp.up_proj.weight
+      value: 0.195161
+    - filter: language_model.model.layers.9.mlp.up_proj.weight
+      value: 0.192881
+    - filter: language_model.model.layers.12.mlp.up_proj.weight
+      value: 0.194623
+    - filter: language_model.model.layers.13.mlp.up_proj.weight
+      value: 0.194623
+    - filter: language_model.model.layers.28.mlp.down_proj.weight
+      value: 0.192328
+    - filter: language_model.model.layers.8.mlp.up_proj.weight
+      value: 0.187722
+    - filter: language_model.model.layers.5.mlp.up_proj.weight
+      value: 0.182478
+    - filter: language_model.model.layers.11.mlp.up_proj.weight
+      value: 0.195844
+    - filter: language_model.model.layers.15.mlp.down_proj.weight
+      value: 0.202757
+    - filter: language_model.model.layers.27.mlp.down_proj.weight
+      value: 0.193527
+    - filter: language_model.model.layers.3.mlp.up_proj.weight
+      value: 0.179504
+    - filter: language_model.model.layers.14.mlp.down_proj.weight
+      value: 0.199945
+    - filter: language_model.model.layers.16.mlp.down_proj.weight
+      value: 0.206862
+    - filter: language_model.model.layers.2.mlp.up_proj.weight
+      value: 0.188289
+    - filter: language_model.model.layers.4.mlp.up_proj.weight
+      value: 0.181225
+    - filter: language_model.model.layers.6.mlp.up_proj.weight
+      value: 0.18594
+    - filter: language_model.model.layers.7.mlp.up_proj.weight
+      value: 0.190633
+    - filter: language_model.model.layers.26.mlp.down_proj.weight
+      value: 0.197091
+    - filter: language_model.model.layers.0.mlp.up_proj.weight
+      value: 0.191835
+    - filter: language_model.model.layers.8.mlp.down_proj.weight
+      value: 0.19655
+    - filter: language_model.model.layers.9.mlp.down_proj.weight
+      value: 0.198898
+    - filter: language_model.model.layers.1.mlp.up_proj.weight
+      value: 0.193633
+    - filter: language_model.model.layers.17.mlp.down_proj.weight
+      value: 0.212397
+    - filter: language_model.model.layers.10.mlp.down_proj.weight
+      value: 0.204918
+    - filter: language_model.model.layers.12.mlp.down_proj.weight
+      value: 0.202561
+    - filter: language_model.model.layers.13.mlp.down_proj.weight
+      value: 0.204918
+    - filter: language_model.model.layers.2.mlp.down_proj.weight
+      value: 0.195448
+    - filter: language_model.model.layers.21.mlp.down_proj.weight
+      value: 0.202561
+    - filter: language_model.model.layers.22.mlp.down_proj.weight
+      value: 0.202561
+    - filter: language_model.model.layers.25.mlp.down_proj.weight
+      value: 0.202561
+    - filter: language_model.model.layers.3.mlp.down_proj.weight
+      value: 0.190673
+    - filter: language_model.model.layers.23.mlp.down_proj.weight
+      value: 0.206783
+    - filter: language_model.model.layers.5.mlp.down_proj.weight
+      value: 0.194886
+    - filter: language_model.model.layers.18.mlp.down_proj.weight
+      value: 0.213388
+    - filter: language_model.model.layers.20.mlp.down_proj.weight
+      value: 0.208664
+    - filter: language_model.model.layers.24.mlp.down_proj.weight
+      value: 0.206291
+    - filter: language_model.model.layers.11.mlp.down_proj.weight
+      value: 0.205793
+    - filter: language_model.model.layers.19.mlp.down_proj.weight
+      value: 0.215302
+    - filter: language_model.model.layers.4.mlp.down_proj.weight
+      value: 0.193743
+    - filter: language_model.model.layers.6.mlp.down_proj.weight
+      value: 0.196167
+    - filter: language_model.model.layers.7.mlp.down_proj.weight
+      value: 0.200995
+    - filter: language_model.model.layers.1.mlp.down_proj.weight
+      value: 0.202878
+    - filter: language_model.model.layers.0.mlp.down_proj.weight
+      value: 0.216568
+    - filter: language_model.model.layers.39.self_attn.o_proj.weight
+      value: 0.125786
+    - filter: language_model.model.layers.38.self_attn.o_proj.weight
+      value: 0.168001
+    - filter: language_model.model.layers.39.self_attn.q_proj.weight
+      value: 0.104322
+    - filter: language_model.model.layers.32.self_attn.q_proj.weight
+      value: 0.142943
+    - filter: language_model.model.layers.35.self_attn.q_proj.weight
+      value: 0.135335
+    - filter: language_model.model.layers.27.self_attn.q_proj.weight
+      value: 0.155682
+    - filter: language_model.model.layers.28.self_attn.q_proj.weight
+      value: 0.160908
+    - filter: language_model.model.layers.36.self_attn.o_proj.weight
+      value: 0.169581
+    - filter: language_model.model.layers.30.self_attn.q_proj.weight
+      value: 0.153355
+    - filter: language_model.model.layers.33.self_attn.q_proj.weight
+      value: 0.14809
+    - filter: language_model.model.layers.37.self_attn.o_proj.weight
+      value: 0.181167
+    - filter: language_model.model.layers.29.self_attn.q_proj.weight
+      value: 0.151579
+    - filter: language_model.model.layers.35.self_attn.o_proj.weight
+      value: 0.172772
+    - filter: language_model.model.layers.5.self_attn.q_proj.weight
+      value: 0.20046
+    - filter: language_model.model.layers.7.self_attn.q_proj.weight
+      value: 0.19533
+    - filter: language_model.model.layers.36.self_attn.q_proj.weight
+      value: 0.131309
+    - filter: language_model.model.layers.6.self_attn.q_proj.weight
+      value: 0.208596
+    - filter: language_model.model.layers.25.self_attn.q_proj.weight
+      value: 0.169983
+    - filter: language_model.model.layers.31.self_attn.q_proj.weight
+      value: 0.159312
+    - filter: language_model.model.layers.2.self_attn.q_proj.weight
+      value: 0.226469
+    - filter: language_model.model.layers.22.self_attn.q_proj.weight
+      value: 0.173013
+    - filter: language_model.model.layers.26.self_attn.q_proj.weight
+      value: 0.169458
+    - filter: language_model.model.layers.34.self_attn.q_proj.weight
+      value: 0.138944
+    - filter: language_model.model.layers.38.self_attn.q_proj.weight
+      value: 0.120914
+    - filter: language_model.model.layers.4.self_attn.q_proj.weight
+      value: 0.206842
+    - filter: language_model.model.layers.3.self_attn.q_proj.weight
+      value: 0.218577
+    - filter: language_model.model.layers.24.self_attn.q_proj.weight
+      value: 0.17765
+    - filter: language_model.model.layers.13.self_attn.q_proj.weight
+      value: 0.19292
+    - filter: language_model.model.layers.32.self_attn.o_proj.weight
+      value: 0.167494
+    - filter: language_model.model.layers.8.self_attn.q_proj.weight
+      value: 0.203614
+    - filter: language_model.model.layers.9.self_attn.q_proj.weight
+      value: 0.201451
+    - filter: language_model.model.layers.33.self_attn.o_proj.weight
+      value: 0.162669
+    - filter: language_model.model.layers.34.self_attn.o_proj.weight
+      value: 0.190237
+    - filter: language_model.model.layers.21.self_attn.q_proj.weight
+      value: 0.191609
+    - filter: language_model.model.layers.12.self_attn.q_proj.weight
+      value: 0.196632
+    - filter: language_model.model.layers.23.self_attn.q_proj.weight
+      value: 0.180131
+    - filter: language_model.model.layers.11.self_attn.q_proj.weight
+      value: 0.207093
+    - filter: language_model.model.layers.37.self_attn.q_proj.weight
+      value: 0.132486
+    - filter: language_model.model.layers.10.self_attn.q_proj.weight
+      value: 0.209974
+    - filter: language_model.model.layers.30.self_attn.o_proj.weight
+      value: 0.159707
+    - filter: language_model.model.layers.14.self_attn.q_proj.weight
+      value: 0.202298
+    - filter: language_model.model.layers.15.self_attn.q_proj.weight
+      value: 0.204133
+    - filter: language_model.model.layers.20.self_attn.q_proj.weight
+      value: 0.196769
+    - filter: language_model.model.layers.18.self_attn.q_proj.weight
+      value: 0.212189
+    - filter: language_model.model.layers.19.self_attn.q_proj.weight
+      value: 0.196912
+    - filter: language_model.model.layers.1.self_attn.q_proj.weight
+      value: 0.257896
+    - filter: language_model.model.layers.16.self_attn.q_proj.weight
+      value: 0.212969
+    - filter: language_model.model.layers.29.self_attn.o_proj.weight
+      value: 0.174884
+    - filter: language_model.model.layers.31.self_attn.o_proj.weight
+      value: 0.17181
+    - filter: language_model.model.layers.17.self_attn.q_proj.weight
+      value: 0.211513
+    - filter: language_model.model.layers.28.self_attn.o_proj.weight
+      value: 0.166682
+    - filter: language_model.model.layers.27.self_attn.o_proj.weight
+      value: 0.168077
+    - filter: language_model.model.layers.26.self_attn.o_proj.weight
+      value: 0.173479
+    - filter: language_model.model.layers.0.self_attn.q_proj.weight
+      value: 0.257296
+    - filter: language_model.model.layers.25.self_attn.o_proj.weight
+      value: 0.186181
+    - filter: language_model.model.layers.22.self_attn.o_proj.weight
+      value: 0.198762
+    - filter: language_model.model.layers.24.self_attn.o_proj.weight
+      value: 0.196575
+    - filter: language_model.model.layers.23.self_attn.o_proj.weight
+      value: 0.196065
+    - filter: language_model.model.layers.21.self_attn.o_proj.weight
+      value: 0.204071
+    - filter: language_model.model.layers.2.self_attn.o_proj.weight
+      value: 0.172557
+    - filter: language_model.model.layers.3.self_attn.o_proj.weight
+      value: 0.158307
+    - filter: language_model.model.layers.4.self_attn.o_proj.weight
+      value: 0.181612
+    - filter: language_model.model.layers.20.self_attn.o_proj.weight
+      value: 0.21766
+    - filter: language_model.model.layers.5.self_attn.o_proj.weight
+      value: 0.17041
+    - filter: language_model.model.layers.19.self_attn.o_proj.weight
+      value: 0.233288
+    - filter: language_model.model.layers.0.self_attn.o_proj.weight
+      value: 0.126586
+    - filter: language_model.model.layers.6.self_attn.o_proj.weight
+      value: 0.187595
+    - filter: language_model.model.layers.7.self_attn.o_proj.weight
+      value: 0.20661
+    - filter: language_model.model.layers.1.self_attn.o_proj.weight
+      value: 0.17396
+    - filter: language_model.model.layers.14.self_attn.o_proj.weight
+      value: 0.216113
+    - filter: language_model.model.layers.8.self_attn.o_proj.weight
+      value: 0.21012
+    - filter: language_model.model.layers.12.self_attn.o_proj.weight
+      value: 0.239032
+    - filter: language_model.model.layers.18.self_attn.o_proj.weight
+      value: 0.239032
+    - filter: language_model.model.layers.9.self_attn.o_proj.weight
+      value: 0.234053
+    - filter: language_model.model.layers.13.self_attn.o_proj.weight
+      value: 0.235535
+    - filter: language_model.model.layers.11.self_attn.o_proj.weight
+      value: 0.235188
+    - filter: language_model.model.layers.15.self_attn.o_proj.weight
+      value: 0.240286
+    - filter: language_model.model.layers.38.self_attn.v_proj.weight
+      value: 0.245061
+    - filter: language_model.model.layers.17.self_attn.o_proj.weight
+      value: 0.259885
+    - filter: language_model.model.layers.39.self_attn.v_proj.weight
+      value: 0.174974
+    - filter: language_model.model.layers.10.self_attn.o_proj.weight
+      value: 0.24449
+    - filter: language_model.model.layers.16.self_attn.o_proj.weight
+      value: 0.247054
+    - filter: language_model.model.layers.0.self_attn.k_proj.weight
+      value: 0.263433
+    - filter: language_model.model.layers.37.self_attn.v_proj.weight
+      value: 0.251838
+    - filter: language_model.model.layers.36.self_attn.v_proj.weight
+      value: 0.229054
+    - filter: language_model.model.layers.2.self_attn.k_proj.weight
+      value: 0.272851
+    - filter: language_model.model.layers.34.self_attn.v_proj.weight
+      value: 0.237781
+    - filter: language_model.model.layers.35.self_attn.v_proj.weight
+      value: 0.234624
+    - filter: language_model.model.layers.39.self_attn.k_proj.weight
+      value: 0.131171
+    - filter: language_model.model.layers.4.self_attn.k_proj.weight
+      value: 0.283584
+    - filter: language_model.model.layers.6.self_attn.k_proj.weight
+      value: 0.27769
+    - filter: language_model.model.layers.5.self_attn.k_proj.weight
+      value: 0.271602
+    - filter: language_model.model.layers.7.self_attn.k_proj.weight
+      value: 0.262444
+    - filter: language_model.model.layers.3.self_attn.k_proj.weight
+      value: 0.287982
+    - filter: language_model.model.layers.1.self_attn.k_proj.weight
+      value: 0.322508
+    - filter: language_model.model.layers.8.self_attn.k_proj.weight
+      value: 0.27421
+    - filter: language_model.model.layers.13.self_attn.k_proj.weight
+      value: 0.264598
+    - filter: language_model.model.layers.9.self_attn.k_proj.weight
+      value: 0.270955
+    - filter: language_model.model.layers.28.self_attn.k_proj.weight
+      value: 0.225454
+    - filter: language_model.model.layers.32.self_attn.v_proj.weight
+      value: 0.223712
+    - filter: language_model.model.layers.10.self_attn.k_proj.weight
+      value: 0.280152
+    - filter: language_model.model.layers.11.self_attn.k_proj.weight
+      value: 0.281748
+    - filter: language_model.model.layers.12.self_attn.k_proj.weight
+      value: 0.268835
+    - filter: language_model.model.layers.15.self_attn.k_proj.weight
+      value: 0.278551
+    - filter: language_model.model.layers.30.self_attn.k_proj.weight
+      value: 0.224603
+    - filter: language_model.model.layers.25.self_attn.k_proj.weight
+      value: 0.246963
+    - filter: language_model.model.layers.27.self_attn.k_proj.weight
+      value: 0.233124
+    - filter: language_model.model.layers.35.self_attn.k_proj.weight
+      value: 0.189342
+    - filter: language_model.model.layers.14.self_attn.k_proj.weight
+      value: 0.275182
+    - filter: language_model.model.layers.26.self_attn.k_proj.weight
+      value: 0.239651
+    - filter: language_model.model.layers.29.self_attn.k_proj.weight
+      value: 0.212543
+    - filter: language_model.model.layers.31.self_attn.k_proj.weight
+      value: 0.224937
+    - filter: language_model.model.layers.22.self_attn.k_proj.weight
+      value: 0.247783
+    - filter: language_model.model.layers.32.self_attn.k_proj.weight
+      value: 0.200668
+    - filter: language_model.model.layers.33.self_attn.k_proj.weight
+      value: 0.215173
+    - filter: language_model.model.layers.24.self_attn.k_proj.weight
+      value: 0.24565
+    - filter: language_model.model.layers.29.self_attn.v_proj.weight
+      value: 0.220374
+    - filter: language_model.model.layers.33.self_attn.v_proj.weight
+      value: 0.227699
+    - filter: language_model.model.layers.18.self_attn.k_proj.weight
+      value: 0.2765
+    - filter: language_model.model.layers.36.self_attn.k_proj.weight
+      value: 0.17181
+    - filter: language_model.model.layers.16.self_attn.k_proj.weight
+      value: 0.289809
+    - filter: language_model.model.layers.21.self_attn.k_proj.weight
+      value: 0.266098
+    - filter: language_model.model.layers.23.self_attn.k_proj.weight
+      value: 0.246661
+    - filter: language_model.model.layers.20.self_attn.k_proj.weight
+      value: 0.266
+    - filter: language_model.model.layers.31.self_attn.v_proj.weight
+      value: 0.219631
+    - filter: language_model.model.layers.30.self_attn.v_proj.weight
+      value: 0.217368
+    - filter: language_model.model.layers.34.self_attn.k_proj.weight
+      value: 0.171745
+    - filter: language_model.model.layers.17.self_attn.k_proj.weight
+      value: 0.281311
+    - filter: language_model.model.layers.19.self_attn.k_proj.weight
+      value: 0.258162
+    - filter: language_model.model.layers.38.self_attn.k_proj.weight
+      value: 0.147819
+    - filter: language_model.model.layers.28.self_attn.v_proj.weight
+      value: 0.212052
+    - filter: language_model.model.layers.27.self_attn.v_proj.weight
+      value: 0.220392
+    - filter: language_model.model.layers.37.self_attn.k_proj.weight
+      value: 0.177511
+    - filter: language_model.model.layers.26.self_attn.v_proj.weight
+      value: 0.223488
+    - filter: language_model.model.layers.22.self_attn.v_proj.weight
+      value: 0.242903
+    - filter: language_model.model.layers.24.self_attn.v_proj.weight
+      value: 0.235792
+    - filter: language_model.model.layers.25.self_attn.v_proj.weight
+      value: 0.231352
+    - filter: language_model.model.layers.23.self_attn.v_proj.weight
+      value: 0.235188
+    - filter: language_model.model.layers.2.self_attn.v_proj.weight
+      value: 0.168169
+    - filter: language_model.model.layers.5.self_attn.v_proj.weight
+      value: 0.165651
+    - filter: language_model.model.layers.3.self_attn.v_proj.weight
+      value: 0.151622
+    - filter: language_model.model.layers.4.self_attn.v_proj.weight
+      value: 0.187611
+    - filter: language_model.model.layers.20.self_attn.v_proj.weight
+      value: 0.224346
+    - filter: language_model.model.layers.19.self_attn.v_proj.weight
+      value: 0.245325
+    - filter: language_model.model.layers.21.self_attn.v_proj.weight
+      value: 0.22807
+    - filter: language_model.model.layers.6.self_attn.v_proj.weight
+      value: 0.18168
+    - filter: language_model.model.layers.7.self_attn.v_proj.weight
+      value: 0.204672
+    - filter: language_model.model.layers.17.self_attn.v_proj.weight
+      value: 0.243401
+    - filter: language_model.model.layers.8.self_attn.v_proj.weight
+      value: 0.19022
+    - filter: language_model.model.layers.12.self_attn.v_proj.weight
+      value: 0.232988
+    - filter: language_model.model.layers.9.self_attn.v_proj.weight
+      value: 0.230431
+    - filter: language_model.model.layers.14.self_attn.v_proj.weight
+      value: 0.205856
+    - filter: language_model.model.layers.1.self_attn.v_proj.weight
+      value: 0.168527
+    - filter: language_model.model.layers.10.self_attn.v_proj.weight
+      value: 0.228984
+    - filter: language_model.model.layers.13.self_attn.v_proj.weight
+      value: 0.219502
+    - filter: language_model.model.layers.15.self_attn.v_proj.weight
+      value: 0.226354
+    - filter: language_model.model.layers.16.self_attn.v_proj.weight
+      value: 0.23457
+    - filter: language_model.model.layers.18.self_attn.v_proj.weight
+      value: 0.250678
+    - filter: language_model.model.layers.11.self_attn.v_proj.weight
+      value: 0.22923
+    - filter: language_model.model.layers.0.self_attn.v_proj.weight
+      value: 0.134132
+    - filter: language_model.model.layers.34.post_attention_layernorm.weight
+      value: 0.883075
+    - filter: language_model.model.layers.33.post_attention_layernorm.weight
+      value: 0.880867
+    - filter: language_model.model.layers.35.post_attention_layernorm.weight
+      value: 0.888537
+    - filter: language_model.model.layers.38.post_attention_layernorm.weight
+      value: 0.895727
+    - filter: language_model.model.layers.31.post_attention_layernorm.weight
+      value: 0.880202
+    - filter: language_model.model.layers.32.post_attention_layernorm.weight
+      value: 0.882647
+    - filter: language_model.model.layers.36.post_attention_layernorm.weight
+      value: 0.895084
+    - filter: language_model.model.layers.37.post_attention_layernorm.weight
+      value: 0.896463
+    - filter: language_model.model.layers.30.post_attention_layernorm.weight
+      value: 0.884693
+    - filter: language_model.model.layers.29.post_attention_layernorm.weight
+      value: 0.886562
+    - filter: language_model.model.layers.28.post_attention_layernorm.weight
+      value: 0.883506
+    - filter: language_model.model.layers.27.post_attention_layernorm.weight
+      value: 0.885518
+    - filter: language_model.model.layers.30.input_layernorm.weight
+      value: 0.900142
+    - filter: language_model.model.layers.26.post_attention_layernorm.weight
+      value: 0.888298
+    - filter: language_model.model.layers.38.input_layernorm.weight
+      value: 0.918705
+    - filter: language_model.model.layers.25.post_attention_layernorm.weight
+      value: 0.887788
+    - filter: language_model.model.layers.11.post_attention_layernorm.weight
+      value: 0.819136
+    - filter: language_model.model.layers.8.input_layernorm.weight
+      value: 0.827776
+    - filter: language_model.model.layers.12.post_attention_layernorm.weight
+      value: 0.829306
+    - filter: language_model.model.layers.17.post_attention_layernorm.weight
+      value: 0.844334
+    - filter: language_model.model.layers.24.post_attention_layernorm.weight
+      value: 0.88749
+    - filter: language_model.model.layers.23.post_attention_layernorm.weight
+      value: 0.882598
+    - filter: language_model.model.layers.18.post_attention_layernorm.weight
+      value: 0.855345
+    - filter: language_model.model.layers.16.post_attention_layernorm.weight
+      value: 0.842616
+    - filter: language_model.model.layers.27.input_layernorm.weight
+      value: 0.906137
+    - filter: language_model.model.layers.13.post_attention_layernorm.weight
+      value: 0.838716
+    - filter: language_model.model.layers.26.input_layernorm.weight
+      value: 0.901367
+    - filter: language_model.model.layers.10.post_attention_layernorm.weight
+      value: 0.815316
+    - filter: language_model.model.layers.14.input_layernorm.weight
+      value: 0.879509
+    - filter: language_model.model.layers.15.post_attention_layernorm.weight
+      value: 0.843714
+    - filter: language_model.model.layers.19.post_attention_layernorm.weight
+      value: 0.865915
+    - filter: language_model.model.layers.22.post_attention_layernorm.weight
+      value: 0.885566
+    - filter: language_model.model.layers.28.input_layernorm.weight
+      value: 0.902543
+    - filter: language_model.model.layers.5.input_layernorm.weight
+      value: 0.810565
+    - filter: language_model.model.layers.6.post_attention_layernorm.weight
+      value: 0.796305
+    - filter: language_model.model.layers.7.post_attention_layernorm.weight
+      value: 0.806038
+    - filter: language_model.model.layers.9.post_attention_layernorm.weight
+      value: 0.815692
+    - filter: language_model.model.layers.21.input_layernorm.weight
+      value: 0.891671
+    - filter: language_model.model.layers.6.input_layernorm.weight
+      value: 0.811936
+    - filter: language_model.model.layers.7.input_layernorm.weight
+      value: 0.83771
+    - filter: language_model.model.layers.14.post_attention_layernorm.weight
+      value: 0.852636
+    - filter: language_model.model.layers.21.post_attention_layernorm.weight
+      value: 0.886217
+    - filter: language_model.model.layers.20.post_attention_layernorm.weight
+      value: 0.880705
+    - filter: language_model.model.layers.3.input_layernorm.weight
+      value: 0.816342
+    - filter: language_model.model.layers.10.input_layernorm.weight
+      value: 0.863017
+    - filter: language_model.model.layers.8.post_attention_layernorm.weight
+      value: 0.820931
+    - filter: language_model.model.layers.1.input_layernorm.weight
+      value: 0.771411
+    - filter: language_model.model.layers.31.input_layernorm.weight
+      value: 0.910857
+    - filter: language_model.model.layers.37.input_layernorm.weight
+      value: 0.922558
+    - filter: language_model.model.layers.5.post_attention_layernorm.weight
+      value: 0.794374
+    - filter: language_model.model.layers.25.input_layernorm.weight
+      value: 0.9114
+    - filter: language_model.model.layers.35.input_layernorm.weight
+      value: 0.923684
+    - filter: language_model.model.layers.33.input_layernorm.weight
+      value: 0.919435
+    - filter: language_model.model.layers.16.input_layernorm.weight
+      value: 0.892982
+    - filter: language_model.model.layers.32.input_layernorm.weight
+      value: 0.921119
+    - filter: language_model.model.layers.4.post_attention_layernorm.weight
+      value: 0.784228
+    - filter: language_model.model.layers.9.input_layernorm.weight
+      value: 0.868681
+    - filter: language_model.model.layers.11.input_layernorm.weight
+      value: 0.884768
+    - filter: language_model.model.layers.36.input_layernorm.weight
+      value: 0.927183
+    - filter: language_model.model.layers.20.input_layernorm.weight
+      value: 0.912409
+    - filter: language_model.model.layers.29.input_layernorm.weight
+      value: 0.92094
+    - filter: language_model.model.layers.34.input_layernorm.weight
+      value: 0.924849
+    - filter: language_model.model.layers.15.input_layernorm.weight
+      value: 0.896905
+    - filter: language_model.model.layers.12.input_layernorm.weight
+      value: 0.891217
+    - filter: language_model.model.layers.13.input_layernorm.weight
+      value: 0.8992
+    - filter: language_model.model.layers.23.input_layernorm.weight
+      value: 0.913613
+    - filter: language_model.model.layers.18.input_layernorm.weight
+      value: 0.906696
+    - filter: language_model.model.layers.24.input_layernorm.weight
+      value: 0.91877
+    - filter: language_model.model.layers.3.post_attention_layernorm.weight
+      value: 0.786238
+    - filter: language_model.model.layers.2.post_attention_layernorm.weight
+      value: 0.777384
+    - filter: language_model.model.layers.39.input_layernorm.weight
+      value: 0.942583
+    - filter: language_model.model.layers.4.input_layernorm.weight
+      value: 0.83397
+    - filter: language_model.model.layers.22.input_layernorm.weight
+      value: 0.916358
+    - filter: language_model.model.layers.17.input_layernorm.weight
+      value: 0.91935
+    - filter: language_model.model.layers.39.post_attention_layernorm.weight
+      value: 0.955214
+    - filter: language_model.model.layers.19.input_layernorm.weight
+      value: 0.932588
+    - filter: language_model.model.layers.0.post_attention_layernorm.weight
+      value: 0.709289
+    - filter: language_model.model.layers.1.post_attention_layernorm.weight
+      value: 0.750784
+    - filter: language_model.model.layers.2.input_layernorm.weight
+      value: 0.851994
+    - filter: language_model.model.layers.0.input_layernorm.weight
+      value: 0.736801
+    - value: 0.0
+tokenizer:
+  source: base
+chat_template: pixtral12b

output.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87d9e3a07c8a729696eb62746361eaa5327f6c6f1f095f6b771a94896616c73c
+size 13701532624

preprocessor_config.json

@@ -0,0 +1,27 @@
+{
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "PixtralImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "patch_size": {
+    "height": 16,
+    "width": 16
+  },
+  "processor_class": "PixtralProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "longest_edge": 1024
+  }
+}

processor_config.json

@@ -0,0 +1,7 @@
+{
+  "image_break_token": "[IMG_BREAK]",
+  "image_end_token": "[IMG_END]",
+  "image_token": "[IMG]",
+  "patch_size": 16,
+  "processor_class": "PixtralProcessor"
+}

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84f33e6f52b2833e8cc17229af8eea363f640a898f19a48184a2c7f6f5a88337
+size 17077329