mradermacher commited on
Commit
a660d27
1 Parent(s): a75a411

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +81 -0
README.md CHANGED
@@ -1,6 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  <!-- ### quantize_version: 2 -->
2
  <!-- ### output_tensor_quantised: 1 -->
3
  <!-- ### convert_type: hf -->
4
  <!-- ### vocab_type: -->
5
  <!-- ### tags: nicoboss -->
6
  weighted/imatrix quants of https://huggingface.co/EVA-UNIT-01/EVA-Qwen2.5-14B-v0.0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: EVA-UNIT-01/EVA-Qwen2.5-14B-v0.0
3
+ datasets:
4
+ - AuriAetherwiing/Allura
5
+ - anthracite-org/kalo-opus-instruct-22k-no-refusal
6
+ - Nopm/Opus_WritingStruct
7
+ - Gryphe/Sonnet3.5-SlimOrcaDedupCleaned
8
+ - Gryphe/Sonnet3.5-Charcard-Roleplay
9
+ - Gryphe/ChatGPT-4o-Writing-Prompts
10
+ - AuriAetherwiing/shortstories_synthlabels
11
+ - Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
12
+ - Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
13
+ - nothingiisreal/Reddit-Dirty-And-WritingPrompts
14
+ language:
15
+ - en
16
+ library_name: transformers
17
+ license: apache-2.0
18
+ quantized_by: mradermacher
19
+ ---
20
+ ## About
21
+
22
  <!-- ### quantize_version: 2 -->
23
  <!-- ### output_tensor_quantised: 1 -->
24
  <!-- ### convert_type: hf -->
25
  <!-- ### vocab_type: -->
26
  <!-- ### tags: nicoboss -->
27
  weighted/imatrix quants of https://huggingface.co/EVA-UNIT-01/EVA-Qwen2.5-14B-v0.0
28
+
29
+ <!-- provided-files -->
30
+ static quants are available at https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-GGUF
31
+ ## Usage
32
+
33
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
34
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
35
+ more details, including on how to concatenate multi-part files.
36
+
37
+ ## Provided Quants
38
+
39
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
40
+
41
+ | Link | Type | Size/GB | Notes |
42
+ |:-----|:-----|--------:|:------|
43
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ1_S.gguf) | i1-IQ1_S | 3.7 | for the desperate |
44
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ1_M.gguf) | i1-IQ1_M | 4.0 | mostly desperate |
45
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 4.4 | |
46
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ2_XS.gguf) | i1-IQ2_XS | 4.8 | |
47
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ2_S.gguf) | i1-IQ2_S | 5.1 | |
48
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ2_M.gguf) | i1-IQ2_M | 5.5 | |
49
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q2_K.gguf) | i1-Q2_K | 5.9 | IQ3_XXS probably better |
50
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 6.0 | lower quality |
51
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ3_XS.gguf) | i1-IQ3_XS | 6.5 | |
52
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q3_K_S.gguf) | i1-Q3_K_S | 6.8 | IQ3_XS probably better |
53
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ3_S.gguf) | i1-IQ3_S | 6.8 | beats Q3_K* |
54
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ3_M.gguf) | i1-IQ3_M | 7.0 | |
55
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q3_K_M.gguf) | i1-Q3_K_M | 7.4 | IQ3_S probably better |
56
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q3_K_L.gguf) | i1-Q3_K_L | 8.0 | IQ3_M probably better |
57
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-IQ4_XS.gguf) | i1-IQ4_XS | 8.2 | |
58
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 8.6 | fast on arm, low quality |
59
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 8.6 | fast on arm+i8mm, low quality |
60
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 8.6 | fast on arm+sve, low quality |
61
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q4_0.gguf) | i1-Q4_0 | 8.6 | fast, low quality |
62
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q4_K_S.gguf) | i1-Q4_K_S | 8.7 | optimal size/speed/quality |
63
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q4_K_M.gguf) | i1-Q4_K_M | 9.1 | fast, recommended |
64
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q5_K_S.gguf) | i1-Q5_K_S | 10.4 | |
65
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q5_K_M.gguf) | i1-Q5_K_M | 10.6 | |
66
+ | [GGUF](https://huggingface.co/mradermacher/EVA-Qwen2.5-14B-v0.0-i1-GGUF/resolve/main/EVA-Qwen2.5-14B-v0.0.i1-Q6_K.gguf) | i1-Q6_K | 12.2 | practically like static Q6_K |
67
+
68
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
69
+ types (lower is better):
70
+
71
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
72
+
73
+ And here are Artefact2's thoughts on the matter:
74
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
75
+
76
+ ## FAQ / Model Request
77
+
78
+ See https://huggingface.co/mradermacher/model_requests for some answers to
79
+ questions you might have and/or if you want some other model quantized.
80
+
81
+ ## Thanks
82
+
83
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
84
+ me use its servers and providing upgrades to my workstation to enable
85
+ this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.
86
+
87
+ <!-- end -->