Add files using upload-large-folder tool

config.json

@@ -0,0 +1,69 @@
+{
+  "_name_or_path": "MarsupialAI/Monstral-123B-v2",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 12288,
+  "initializer_range": 0.02,
+  "intermediate_size": 28672,
+  "max_position_embeddings": 131072,
+  "model_type": "mistral",
+  "num_attention_heads": 96,
+  "num_hidden_layers": 88,
+  "num_key_value_heads": 8,
+  "quantization_config": {
+    "config_groups": {
+      "group_0": {
+        "input_activations": {
+          "actorder": null,
+          "block_structure": null,
+          "dynamic": true,
+          "group_size": null,
+          "num_bits": 8,
+          "observer": null,
+          "observer_kwargs": {},
+          "strategy": "token",
+          "symmetric": true,
+          "type": "float"
+        },
+        "output_activations": null,
+        "targets": [
+          "Linear"
+        ],
+        "weights": {
+          "actorder": null,
+          "block_structure": null,
+          "dynamic": false,
+          "group_size": null,
+          "num_bits": 8,
+          "observer": "minmax",
+          "observer_kwargs": {},
+          "strategy": "channel",
+          "symmetric": true,
+          "type": "float"
+        }
+      }
+    },
+    "format": "float-quantized",
+    "global_compression_ratio": 1.4645315540129722,
+    "ignore": [
+      "lm_head"
+    ],
+    "kv_cache_scheme": null,
+    "quant_method": "compressed-tensors",
+    "quantization_status": "compressed"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.46.3",
+  "use_cache": true,
+  "vocab_size": 32768
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.46.3"
+}

model-00001-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8a8941db2c557dbd5ae84f64e49d49ef2eba34f7d20f310dbdfbd31462f7ebc
+size 4958398104

model-00002-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d837eade7acf71492a2fe2315a3250ed5110211159a7e9d35f00d2806fc6039
+size 4832680456

model-00003-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:314155b8a72732502f1d0ca296703add6c1c5a44a0d5d09dabb21859d26363a2
+size 4857866272

model-00004-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61d17405a3ebf36b80d79eafba4b60586fcc34b2a44ec055ba14aa73f1016a19
+size 4832680520

model-00005-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dde6fe9bc938cf36ce7d4cc245dd6772ba68d7df4b50b8436c40c72f94185061
+size 4857866328

model-00006-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1bb589ad04b445ae8ac5a9c66a7f62ee9c811321988482b59248dad535e611b
+size 4832680520

model-00007-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:966b9074f799bf48c0d3c5a349defa4fb001c5db346cca311876b8a295410b60
+size 4857866328

model-00008-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:713845122093586dafc4ed1ce1529379e6685ff20cc24a1be81b78a89157532c
+size 4832680520

model-00009-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29ef9ff1ffac2916e9e83d0c0ececc0e1b29ec9397e324f1a6b37ddb2baf805e
+size 4857866328

model-00010-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5ca401ad125a4276e02b06197c43e984777caa593f055164e8fcced930a4139
+size 4832680520

model-00011-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b578228e33400421386e9af57c384c86caf268275aa90f48e793d3eaa78caefa
+size 4857866328

model-00012-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:efbc1d7a19e8eb811ff4233284053a30f3e5c4870918692337e2580aac909ca1
+size 4832680520

model-00013-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db96aa8af0a3f7eee4c2d33948a55f3cc5104b33905f47b5c5894f3533962c90
+size 4857866328

model-00014-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c50710a0a3191cf5dffc73d7dcc0f237e785035db658ae87ab3d7dfa4b8770a
+size 4832680520

model-00015-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7a076c9b3c5e545ff5a448e504f4a22a8ad0820f8bcfa0d3c8ba8da724e26b3
+size 4857866328

model-00016-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:694abe5eafa86497ab54785b9b1e782c81fcb48c995e95e873a163a8d3094fde
+size 4832680520

model-00017-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dfffc82426799b38782b70838dbe63c2cbd8920b3cdb56ad1c6d8f42f58c67a9
+size 4857866328

model-00018-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd3e73901d50e6033d8dae48609b2c05744f83ad953e4ef8291793c7fc20c3bb
+size 4832680520

model-00019-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84ff90580a4dbf5128b3fbaea9f135fd3cc74f858da7c0722a05036eb7e85948
+size 4857866328

model-00020-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4ac936c880f404d05ed4466e1abd9071ae30c8b5e7546ff52445eda200ceb08
+size 4832680520

model-00021-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:800808a4d3ef0329371128acc2b90940778fbc722bd54cf5deb51ae1f449ff5b
+size 4857866328

model-00022-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7896906bb8c1322ee471b65edec26633946f57a4eb619e6c7a00d4114a6502b
+size 4832680520

model-00023-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a83529141cab842b5f875cba029e236f6390a939d65dcb5b4f00bffe3ad03896
+size 4857866328

model-00024-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f296fec3c2c217ddd4d1110f0a60885e1ae58991367f6c89251fbff2c69a8552
+size 4832680520

model-00025-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee169de644517e9677f8c69f1c9ea2c96fd6fb099afc08e9614e94b2b35db2c4
+size 4857866328

model-00026-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d939ea22d73e95a67b96769acb1ea7e5171f94d083fc3bba49b3e8fcdf0332d6
+size 2189695040

recipe.yaml

@@ -0,0 +1,6 @@
+DEFAULT_stage:
+  DEFAULT_modifiers:
+    QuantizationModifier:
+      ignore: [lm_head]
+      targets: [Linear]
+      scheme: FP8_DYNAMIC

special_tokens_map.json

@@ -0,0 +1,23 @@
+{
+  "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
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}