new dispatch again + correct dtype

config.json

@@ -1,20 +1,10 @@
 {
-  "_name_or_path": "/extra_disk_1/dkuznedelev/aqlm/Llama-2-70b-AQLM-2Bit-1x16-hf",
-  "aqlm": {
-    "in_group_size": 8,
-    "nbits_per_codebook": 16,
-    "num_codebooks": 1,
-    "out_group_size": 1
-  },
+  "_name_or_path": "BlackSamorez/Llama-2-70b-AQLM-2Bit-1x16-hf",
   "architectures": [
     "LlamaForCausalLM"
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
-  "auto_map": {
-    "AutoConfig": "configuration_llama_aqlm.LlamaConfig",
-    "AutoModelForCausalLM": "modeling_llama_aqlm.LlamaForCausalLM"
-  },
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
@@ -22,17 +12,28 @@
   "initializer_range": 0.02,
   "intermediate_size": 28672,
   "max_position_embeddings": 4096,
-  "model_type": "llama_aqlm",
+  "model_type": "llama",
   "num_attention_heads": 64,
   "num_hidden_layers": 80,
   "num_key_value_heads": 8,
   "pretraining_tp": 1,
+  "quantization_config": {
+    "in_group_size": 8,
+    "linear_weights_not_to_quantize": [
+      "model.embed_tokens.weight",
+      "lm_head.weight"
+    ],
+    "nbits_per_codebook": 16,
+    "num_codebooks": 1,
+    "out_group_size": 1,
+    "quant_method": "aqlm"
+  },
   "rms_norm_eps": 1e-05,
   "rope_scaling": null,
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
-  "torch_dtype": "float32",
-  "transformers_version": "4.37.2",
+  "torch_dtype": "float16",
+  "transformers_version": "4.38.0",
   "use_cache": true,
   "vocab_size": 32000
 }

generation_config.json

@@ -2,5 +2,5 @@
   "_from_model_config": true,
   "bos_token_id": 1,
   "eos_token_id": 2,
-  "transformers_version": "4.37.2"
+  "transformers_version": "4.38.0"
 }

model-00001-of-00005.safetensors → model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a9dc9c6d4d68cb998a9602f81f240e0fb1ffc333aeabed3746f4af44d154c51e
-size 4989778312
+oid sha256:60e85d108d68e360d0342453e2106d43d67bbc883923b138f9fe7c2f792f7f76
+size 4996376224

model-00002-of-00005.safetensors → model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4ca83217933c3d4aa734853be1ed897bd4f375e48fbb2dee351c15e0b20e7978
-size 4991647616
+oid sha256:63f1e45dca214aa63179040ac183b2a90468a30f1a7e4b6aedd463369bee5688
+size 4991616704

model-00003-of-00005.safetensors → model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c1f1e4dd22423b61d6a3550cc78af488f35a1926aa218a48ddb07879db2da29
-size 4976943808
+oid sha256:1e66eefba897aedfcecec42a2d89286e8e17a364ee6734649471ebc85c5ff51b
+size 4973763520

model-00004-of-00005.safetensors → model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:449da0466ca4bc5eec0a0a557319a27794c8743002a6055dccf3e5e5c3b6498c
-size 4409730992
+oid sha256:d808e60f33b9aac1c9a75e6def7f9762061b6487f9484a3f6ca3af2940f37087
+size 3803068312

model-00005-of-00005.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:34576756203b32b277f651fc65fa3f8f7171e7bb5796ec52bbeb4e5bb4a645af
-size 1048576128