Commit
•
01dd88e
1
Parent(s):
0967bee
uploaded readme
Browse files
README.md
ADDED
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Quantization made by Richard Erkhov.
|
2 |
+
|
3 |
+
[Github](https://github.com/RichardErkhov)
|
4 |
+
|
5 |
+
[Discord](https://discord.gg/pvy7H8DZMG)
|
6 |
+
|
7 |
+
[Request more models](https://github.com/RichardErkhov/quant_request)
|
8 |
+
|
9 |
+
|
10 |
+
llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged - GGUF
|
11 |
+
- Model creator: https://huggingface.co/dhmeltzer/
|
12 |
+
- Original model: https://huggingface.co/dhmeltzer/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged/
|
13 |
+
|
14 |
+
|
15 |
+
| Name | Quant method | Size |
|
16 |
+
| ---- | ---- | ---- |
|
17 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q2_K.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q2_K.gguf) | Q2_K | 2.36GB |
|
18 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ3_XS.gguf) | IQ3_XS | 2.6GB |
|
19 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ3_S.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ3_S.gguf) | IQ3_S | 2.75GB |
|
20 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K_S.gguf) | Q3_K_S | 2.75GB |
|
21 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ3_M.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ3_M.gguf) | IQ3_M | 2.9GB |
|
22 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K.gguf) | Q3_K | 3.07GB |
|
23 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K_M.gguf) | Q3_K_M | 3.07GB |
|
24 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q3_K_L.gguf) | Q3_K_L | 3.35GB |
|
25 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ4_XS.gguf) | IQ4_XS | 3.4GB |
|
26 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_0.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_0.gguf) | Q4_0 | 3.56GB |
|
27 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.IQ4_NL.gguf) | IQ4_NL | 3.58GB |
|
28 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_K_S.gguf) | Q4_K_S | 3.59GB |
|
29 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_K.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_K.gguf) | Q4_K | 3.8GB |
|
30 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_K_M.gguf) | Q4_K_M | 3.8GB |
|
31 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_1.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q4_1.gguf) | Q4_1 | 3.95GB |
|
32 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_0.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_0.gguf) | Q5_0 | 4.33GB |
|
33 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_K_S.gguf) | Q5_K_S | 4.33GB |
|
34 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_K.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_K.gguf) | Q5_K | 4.45GB |
|
35 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_K_M.gguf) | Q5_K_M | 4.45GB |
|
36 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_1.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q5_1.gguf) | Q5_1 | 4.72GB |
|
37 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q6_K.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q6_K.gguf) | Q6_K | 5.15GB |
|
38 |
+
| [llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q8_0.gguf](https://huggingface.co/RichardErkhov/dhmeltzer_-_llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged-gguf/blob/main/llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged.Q8_0.gguf) | Q8_0 | 6.67GB |
|
39 |
+
|
40 |
+
|
41 |
+
|
42 |
+
|
43 |
+
Original model description:
|
44 |
+
|
45 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
46 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_dhmeltzer__llama-7b-SFT_eli5_wiki65k_1024_r_64_alpha_16_merged)
|
47 |
+
|
48 |
+
| Metric | Value |
|
49 |
+
|-----------------------|---------------------------|
|
50 |
+
| Avg. | 43.96 |
|
51 |
+
| ARC (25-shot) | 53.75 |
|
52 |
+
| HellaSwag (10-shot) | 78.76 |
|
53 |
+
| MMLU (5-shot) | 46.02 |
|
54 |
+
| TruthfulQA (0-shot) | 43.31 |
|
55 |
+
| Winogrande (5-shot) | 73.48 |
|
56 |
+
| GSM8K (5-shot) | 4.7 |
|
57 |
+
| DROP (3-shot) | 7.72 |
|
58 |
+
|
59 |
+
|