Upload folder using huggingface_hub

README.md

@@ -43,8 +43,5 @@ python /home/alpha/AI/exllamav2/convert.py --in_dir /home/alpha/FastModels/v8/v8
 
 Second pass:
 ```
-python /home/alpha/AI/exllamav2/convert.py --in_dir /home/alpha/FastModels/v8/v8 -o /home/alpha/FastModels/scratch -m /home/alpha/FastModels/v8meas.json --cal_dataset /home/alpha/Documents/stories.parquet -l 12288 -r 26 -ml 32768 -mr 8 -ss 4096 -b 3.1 -hb 6 -cf /home/alpha/FastModels/v8-exl2-31bpw-fiction -nr
+python /home/alpha/AI/exllamav2/convert.py --in_dir /home/alpha/FastModels/v8/v8 -o /home/alpha/FastModels/scratch -m /home/alpha/FastModels/v8meas.json --cal_dataset /home/alpha/Documents/stories.parquet -l 12288 -r 26 -ml 32768 -mr 8 -ss 4096 -b 4.0 -hb 6 -cf /home/alpha/FastModels/v8-exl2-4bpw-fiction -nr
 ```
-
-
-

config.json

@@ -4,7 +4,6 @@
     "LlamaForCausalLM"
   ],
   "attention_bias": false,
-  "attention_dropout": 0.0,
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
@@ -23,7 +22,7 @@
   "rope_theta": 5000000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.36.2",
+  "transformers_version": "4.35.2",
   "use_cache": true,
   "vocab_size": 64002
 }

mergekit_config.yml

@@ -1,51 +1,60 @@
-models:
-  - model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
-    # No parameters necessary for base model
-  - model: /home/alpha/Storage/Models/Raw/migtissera_Tess-34B-v1.4
-    parameters:
-      weight: [0.23, 0.125, 0.125, 0.125, 0.125, 0.125]
-      density: 0.59
-  - model: /home/alpha/Models/Raw/Mihaiii_Pallas-0.5
-    parameters:
-      weight: [0.23, 0.125, 0.125, 0.125, 0.125, 0.125]
-      density: 0.59
-  - model: /home/alpha//Storage/Models/Raw/bhenrym14_airoboros-3_1-yi-34b-200k
+base_model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
+dtype: bfloat16
+merge_method: dare_ties
+parameters:
+  int8_mask: 1.0
+slices:
+- sources:
+  - layer_range: [0, 60]
+    model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
+  - layer_range: [0, 60]
+    model: /home/alpha/Storage/Models/Raw/migtissera_Tess-34B-v1.4
+    parameters:
+      density: 0.61
+      weight: [0.22, 0.113, 0.113, 0.113, 0.113, 0.113]
+  - layer_range: [0, 60]
+    model: /home/alpha/Models/Raw/Mihaiii_Pallas-0.5
+    parameters:
+      density: 0.61
+      weight: [0.22, 0.113, 0.113, 0.113, 0.113, 0.113]
+  - layer_range: [0, 60]
+    model: /home/alpha//Storage/Models/Raw/bhenrym14_airoboros-3_1-yi-34b-200k
     parameters:
-      weight: [0.02, 0.106, 0.106, 0.106, 0.106, 0.106]
       density: 0.59
-  - model: /home/alpha/Storage/Models/Raw/jondurbin_bagel-34b-v0.2
-    #Only the SFT in the main merge since the DPO version seems to have no long context ability at all
+      weight: [0.02, 0.081, 0.081, 0.081, 0.081, 0.081]
+  - layer_range: [0, 60]
+    model: /home/alpha/Storage/Models/Raw/jondurbin_bagel-34b-v0.2
     parameters:
-      weight: [0.02, 0.100, 0.100, 0.100, 0.100, 0.100]
       density: 0.4
-  - model: /home/alpha/Storage/Models/Raw/kyujinpy_PlatYi-34B-200k-Q-FastChat
+      weight: [0.02, 0.093, 0.093, 0.093, 0.093, 0.093]
+  - layer_range: [0, 60]
+    model: /home/alpha/Storage/Models/Raw/kyujinpy_PlatYi-34B-200k-Q-FastChat
     parameters:
-      weight: [0.02, 0.100, 0.100, 0.100, 0.100, 0.100]
       density: 0.59
-  #- model: /home/alpha/Storage/Models/Raw/ehartford_dolphin-2.2-yi-34b-200k
-  #  Dolphin 200K seems to be funky according to multiple leaderboards and perplexity tests?
-  #  parameters:
-  #    weight: 0.15
-  #    density: 0.6
-  - model: /home/alpha/Models/Raw/adamo1139_Yi-34B-200K-AEZAKMI-v2
-    parameters:
-      weight: [0.02, 0.110, 0.110, 0.110, 0.110, 0.110]
-      density: 0.59
-  - model: /home/alpha/Storage/Models/Raw/Nous-Capybara-34B
+      weight: [0.02, 0.081, 0.081, 0.081, 0.081, 0.081]
+  - layer_range: [0, 60]
+    model: /home/alpha/Models/Raw/adamo1139_Yi-34B-200K-AEZAKMI-v2
     parameters:
-      weight:  [0.22, 0.126, 0.126, 0.126, 0.126, 0.126]
       density: 0.59
-  - model: /home/alpha/Storage/Models/Raw/4kmerge
-    parameters:
-      weight: [0.02,  0.108, 0.108, 0.108, 0.108, 0.108]
-      density: 0.5
-  - model: /home/alpha/Models/Raw/migtissera_Tess-M-Creative-v1.0
+      weight: [0.02, 0.096, 0.096, 0.096, 0.096, 0.096]
+  - layer_range: [0, 60]
+    model: /home/alpha/Storage/Models/Raw/Nous-Capybara-34B
     parameters:
-      weight: [0.22, 0.100, 0.100, 0.100, 0.100, 0.10]
       density: 0.59
-merge_method: dare_ties
-tokenizer_source: union
-base_model: /home/alpha/Storage/Models/Raw/chargoddard_Yi-34B-200K-Llama
-parameters:
-  int8_mask: true
-dtype: bfloat16
+      weight: [0.21, 0.115, 0.115, 0.115, 0.115, 0.115]
+  - layer_range: [0, 60]
+    model: /home/alpha/FastModels/v8/4kmerge-v2
+    parameters:
+      density: 0.4
+      weight: [0.02, 0.115, 0.115, 0.115, 0.115, 0.115]
+  - layer_range: [0, 60]
+    model: /home/alpha/Models/Raw/migtissera_Tess-M-Creative-v1.0
+    parameters:
+      density: 0.61
+      weight: [0.21, 0.09, 0.09, 0.09, 0.09, 0.09]
+  - layer_range: [0, 60]
+    model: /home/alpha/Models/Raw/TriadParty_deepmoney-34b-200k-base
+    parameters:
+      density: 0.61
+      weight: [0.04, 0.103, 0.103, 0.103, 0.103, 0.103]
+tokenizer_source: union

output-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe7847b6bd38ad0aec4cd13922152747fd2865a39151d395b60bc8d927fbd415
-size 4243474688
+oid sha256:741ec744d2283a49c5bb87fe353ea0b8f9a13104429da8e20232913d44153946
+size 4294169880

output-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a99b7250790c05a7bfcbc65f867ac4e4093d6ec5d12287bd53addf0c938d4b14
-size 4292411376
+oid sha256:512c23152ebd0345ea2dd0c4844ff9b9e63cac2a279f4bfc43acc6f2ba214faf
+size 4256670720

output-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90b2ba416bb84d138a4398ac16629fcbf02c1ed07086604f7570f6ab54a50ccf
-size 4294489168
+oid sha256:ec4036118c1c00b354f01a9b50a1c00272f95b713a26babebfba4783c097c266
+size 4284799408

output-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25b4f037a70aeb3f9a7e392082cc09d5c2b9ba19da336922b5d1173df4c28b52
-size 1437881976
+oid sha256:e2f0dffa49ff18c10fdb02fd1add1f85f60d70a7d4c63bc28ac46e831f043545
+size 1433196720

tokenizer.json

@@ -69,6 +69,12 @@
   "post_processor": {
     "type": "TemplateProcessing",
     "single": [
+      {
+        "SpecialToken": {
+          "id": "<|startoftext|>",
+          "type_id": 0
+        }
+      },
       {
         "Sequence": {
           "id": "A",
@@ -77,12 +83,24 @@
       }
     ],
     "pair": [
+      {
+        "SpecialToken": {
+          "id": "<|startoftext|>",
+          "type_id": 0
+        }
+      },
       {
         "Sequence": {
           "id": "A",
           "type_id": 0
         }
       },
+      {
+        "SpecialToken": {
+          "id": "<|startoftext|>",
+          "type_id": 1
+        }
+      },
       {
         "Sequence": {
           "id": "B",
@@ -90,7 +108,17 @@
         }
       }
     ],
-    "special_tokens": {}
+    "special_tokens": {
+      "<|startoftext|>": {
+        "id": "<|startoftext|>",
+        "ids": [
+          1
+        ],
+        "tokens": [
+          "<|startoftext|>"
+        ]
+      }
+    }
   },
   "decoder": {
     "type": "Sequence",

tokenizer_config.json

@@ -1,6 +1,4 @@
 {
-  "add_bos_token": false,
-  "add_eos_token": false,
   "added_tokens_decoder": {
     "0": {
       "content": "<unk>",