Nabushika commited on
Commit
96fe798
·
verified ·
1 Parent(s): 966b4e6

Upload folder using huggingface_hub

Browse files
README.md ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ ---
4
+ # Join our Discord! https://discord.gg/Nbv9pQ88Xb
5
+ ## Nearly 2500 members strong 💪
6
+ ### Now with more channels! A hub for creatives and makers alike!
7
+ ---
8
+
9
+ [BeaverAI](https://huggingface.co/BeaverAI) proudly presents...
10
+
11
+ *The finetune that made people buy another 3090...*
12
+
13
+ # Behemoth 123B v2.2 🦣
14
+
15
+ > Nothing in the void is foreign to us. The place we go is the place we belong.
16
+
17
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/fLdJM1oTjLpEKJsbl1BB7.png)
18
+
19
+ ## Links
20
+ - Original: https://huggingface.co/TheDrummer/Behemoth-123B-v2.2
21
+ - GGUF: https://huggingface.co/TheDrummer/Behemoth-123B-v2.2-GGUF
22
+ - iMatrix: https://huggingface.co/bartowski/Behemoth-123B-v2.2-GGUF (recommended for smaller quants)
23
+
24
+ ## Description
25
+
26
+ Behemoth v2.x is a finetune of the new Largestral 2411 with system prompt support. Testers have noted that **everything** felt improved.
27
+
28
+ ### Usage
29
+ Testers say this frankenformat maximizes the model's potential: **Metharme** with Mistral's new system tokens
30
+ - `[SYSTEM_PROMPT] <|system|>{{system_message}}[/SYSTEM_PROMPT]<|user|>{{user_message}}<|model|>{{assistant_message}}`
31
+ - `<|system|>[SYSTEM_PROMPT] {{system_message}}[/SYSTEM_PROMPT]<|user|>{{user_message}}<|model|>{{assistant_message}}`
32
+
33
+ *Take note that the opening system tag SHOULD ALWAYS have a leading whitespace after it.*
34
+
35
+ Complete SillyTavern Settings in BeaverAI Club: https://discord.com/channels/1238219753324281886/1309968730301792370/1309968730301792370
36
+
37
+ Mirror: https://rentry.org/cd32disa
38
+
39
+ ### Versions
40
+ - [v2.0](https://huggingface.co/TheDrummer/Behemoth-123B-v2) is equivalent to Behemoth v1.0 (Classic)
41
+ - [v2.1](https://huggingface.co/TheDrummer/Behemoth-123B-v2.1) is equivalent to Behemoth v1.1 (Creative Boost)
42
+ - [v2.2](https://huggingface.co/TheDrummer/Behemoth-123B-v2.2) is an improvement of Behemoth v2.1 (Creative++)
43
+
44
+ ## Special Thanks
45
+
46
+ Thank you to each and everyone who donated/subscribed in [Ko-Fi](https://ko-fi.com/thedrummer) 🙇 I hope to never disappoint!
47
+
48
+ ```
49
+ Toasty Pigeon
50
+ theguywhogamesalot
51
+ Grozi
52
+ F
53
+ Marinara
54
+ Ko-fi Supporter
55
+ Grozi
56
+ Phaelon
57
+ ONTHEREDTEAM
58
+ EvarinSharath'fe(USM-Valor)
59
+ Silva
60
+ Dakkidaze
61
+ AlexTheVP
62
+ Pseudo
63
+ Kistara
64
+ Dr. Fjut
65
+ Grozi 🥈
66
+ KinjiHakari777
67
+ dustywintr
68
+ Syd
69
+ HumbleConsumer
70
+ Syd
71
+ Ko-fi Supporter
72
+ Arkamist
73
+ joe 🥇
74
+ Toad
75
+ Lied
76
+ Konnect
77
+ Kistara
78
+ Grozi 🥉
79
+ SleepDeprived3
80
+ Luigi
81
+ Nestor
82
+ ```
83
+
84
+ https://ko-fi.com/thedrummer/leaderboard
85
+
86
+ ```
87
+ Finetuned by yours truly,
88
+ Drummer
89
+ ```
90
+
91
+ Thank you Gargy for the GPUs!
92
+
93
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/KvyYIIA1zkxQNEdGro007.png)
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "merged/BEHEMOTH-SLERP",
3
+ "architectures": [
4
+ "MistralForCausalLM"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 12288,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 28672,
14
+ "max_position_embeddings": 131072,
15
+ "model_type": "mistral",
16
+ "num_attention_heads": 96,
17
+ "num_hidden_layers": 88,
18
+ "num_key_value_heads": 8,
19
+ "rms_norm_eps": 1e-05,
20
+ "rope_theta": 1000000.0,
21
+ "sliding_window": null,
22
+ "tie_word_embeddings": false,
23
+ "torch_dtype": "bfloat16",
24
+ "transformers_version": "4.46.3",
25
+ "use_cache": true,
26
+ "vocab_size": 32768
27
+ }
measurement.json ADDED
The diff for this file is too large to render. See raw diff
 
output-00001-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39f4c4a8d15e6840564f4bde627830b3e3d9a87b9e4e87b200d1c2129d686ec4
3
+ size 8515542238
output-00002-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:935714664f652262904eac7cb9c9afaf025ba06db80e4dbdc864cf4ac8e14e1d
3
+ size 8516639342
output-00003-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1968700acc0efba5c238f3bf741a411b21d9dd36e801cfdf1383665bacf67717
3
+ size 8551799034
output-00004-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:16a89e24ff95ace9d594e7946b77e47dd92aa1428507b64dc3e42d5e14483769
3
+ size 8558713330
output-00005-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a93fc85e78fc300f3fdb069034da2ab4ee5ac381b8b19a870444d456fcebcf4
3
+ size 7306067232
special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "unk_token": {
17
+ "content": "<unk>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b968b8dc352f42192367337c78ccc61e1eaddc6d641a579372d4f20694beb7a
3
+ size 587562
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff