Add files using upload-large-folder tool

config.json

@@ -0,0 +1,69 @@
+{
+  "_name_or_path": "MarsupialAI/Monstral-123B",
+  "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 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "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:f6e05b301c526b3058f2f17491373c694eba6aa4418155347b3ac9ef1952fc9b
+size 4958398104

model-00002-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88d572f003d854f47ff0f9f3f74a78b3930411eae6131aef03c67ebf0e87f789
+size 4832680456

model-00003-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35a1868c61d6439de3000b5bc2aeed78d40b32cc4bfcb884d48f635b61340236
+size 4857866272

model-00004-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46f0e89a6654a3d6f21eba660904f9b7ce801d3fbd2f130b24a55565b2162134
+size 4832680520

model-00005-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8910a067ab7e48332e9edc6cef2e3e14def2f922601dacee9e261c2521b60a0d
+size 4857866328

model-00006-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d52723de89cea163bebe9b66a6c34224e144723b57457ff3daaf87879d12c97
+size 4832680520

model-00007-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54eff59c800132b02267cfe120a963c6d6ac1adb177b33dda316062f8a489568
+size 4857866328

model-00008-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9db1c4003fdfb2c0d49bad340b7c8476223843ff1d9ac604b7c091efb59e135
+size 4832680520

model-00009-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:effd8789f53b7ffe174509ceb479211a05c79cbd982c3ea075fb3ea70d419444
+size 4857866328

model-00010-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d4eac075f12af09a239d5ff49eaf43078a9b3b213ee29756f1c613242f06614
+size 4832680520

model-00011-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0baf419a435872a0b51a37b40090d6a34588d3d4ca72f7f3b7dc15f3622176f
+size 4857866328

model-00012-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6910bc6fa74552f33265fe0d08a299c648e5757bad8558481d301577ff60f989
+size 4832680520

model-00013-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78fd4d79d02069ad0429a68c66d3a8cbb76fde5c57c074acc40e8aa0f881ceec
+size 4857866328

model-00014-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:453b0bd2ff48f932e1df68aa305df0b5146bfd6fcaa1898794502ee758d5f44b
+size 4832680520

model-00015-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddc8a767f4b7739f70a29ac9182d18fb738f3e2a569b5b6efd7474eaf34103ca
+size 4857866328

model-00016-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:855e435eb17e9a31bc839e6119e4824ab87d43adc7d667720de62c873c6bc615
+size 4832680520

model-00017-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97205afb6dfe38cb4df3435406cbd8a22fb6fec74d7734c38683d603f78443cb
+size 4857866328

model-00018-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a7e4bc9a3e7148c6315adf044c82de6668c1a1d5a29f38362b4804f15f0b986
+size 4832680520

model-00019-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2427dc9cd9fba257526aa6dbe99757c89b019682ee0b082930557263f3130598
+size 4857866328

model-00020-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15133da40e50056b1e8d4bcae65659c643b70354337f803e642b1fff4eb9bdc2
+size 4832680520

model-00021-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38ca180b8366da2737f12a563b088749eb554ac7d986f09562f5eec79ce4a3b8
+size 4857866328

model-00022-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24a1494047244a4a37e12e1ddbd81e555620bfbc9b975b3083426c376d1755ae
+size 4832680520

model-00023-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b586080da70f423bfe8bcb88efc28622bff10690b8c4649c449691ef56fa6040
+size 4857866328

model-00024-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fdabe88f3c80dc7b3b588f66ae1cdc7b212111653532fbb204922742db3e3493
+size 4832680520

model-00025-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4929a5020e051f16d51e1affc28a00023a54f21c7cd39c90cb36dcc504e278fc
+size 4857866328

model-00026-of-00026.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:175a4cc7e74cee318e0401201afb70b7c2e449103ce4b44a1069c9efc872803f
+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
+  }
+}