Conversational model based on Llama 2 that performs competitively on various benchmarks.
11.2K Pulls Updated 6 months ago
70b-v0.1
39GB
70b-v0.1-q4_0
39GB
70b-v0.1-q4_1
43GB
70b-v0.1-q5_0
47GB
70b-v0.1-q5_1
52GB
70b-v0.1-q8_0
73GB
70b-v0.1-q2_K
29GB
70b-v0.1-q3_K_S
30GB
70b-v0.1-q3_K_M
33GB
70b-v0.1-q3_K_L
36GB
70b-v0.1-q4_K_S
39GB
70b-v0.1-q4_K_M
41GB
70b-v0.1-q5_K_S
47GB
70b-v0.1-q6_K
57GB
70b-v0.1-fp16
138GB
13b-v0.1
7.4GB
13b-v0.2
7.4GB
13b-v0.1-q4_0
7.4GB
13b-v0.1-q4_1
8.2GB
13b-v0.1-q5_0
9.0GB
13b-v0.1-q5_1
9.8GB
13b-v0.1-q8_0
14GB
13b-v0.1-q2_K
5.4GB
13b-v0.1-q3_K_S
5.7GB
13b-v0.1-q3_K_M
6.3GB
13b-v0.1-q3_K_L
6.9GB
13b-v0.1-q4_K_S
7.4GB
13b-v0.1-q4_K_M
7.9GB
13b-v0.1-q5_K_S
9.0GB
13b-v0.1-q5_K_M
9.2GB
13b-v0.1-q6_K
11GB
13b-v0.1-fp16
26GB
13b-v0.2-q4_0
7.4GB
13b-v0.2-q4_1
8.2GB
13b-v0.2-q5_0
9.0GB
13b-v0.2-q5_1
9.8GB
13b-v0.2-q8_0
14GB
13b-v0.2-q2_K
5.4GB
13b-v0.2-q3_K_S
5.7GB
13b-v0.2-q3_K_M
6.3GB
13b-v0.2-q3_K_L
6.9GB
13b-v0.2-q4_K_S
7.4GB
13b-v0.2-q4_K_M
7.9GB
13b-v0.2-q5_K_S
9.0GB
13b-v0.2-q5_K_M
9.2GB
13b-v0.2-q6_K
11GB
13b-v0.2-fp16
26GB
7b-v0.1
3.8GB
7b-v0.2
3.8GB
7b-v0.1-q4_0
3.8GB
7b-v0.1-q4_1
4.2GB
7b-v0.1-q5_0
4.7GB
7b-v0.1-q5_1
5.1GB
7b-v0.1-q8_0
7.2GB
7b-v0.1-q2_K
2.8GB
7b-v0.1-q3_K_S
2.9GB
7b-v0.1-q3_K_M
3.3GB
7b-v0.1-q3_K_L
3.6GB
7b-v0.1-q4_K_S
3.9GB
7b-v0.1-q4_K_M
4.1GB
7b-v0.1-q5_K_S
4.7GB
7b-v0.1-q5_K_M
4.8GB
7b-v0.1-q6_K
5.5GB
7b-v0.1-fp16
13GB
7b-v0.2-q4_0
3.8GB
7b-v0.2-q4_1
4.2GB
7b-v0.2-q5_0
4.7GB
7b-v0.2-q8_0
7.2GB
7b-v0.2-q2_K
2.8GB
7b-v0.2-q3_K_S
2.9GB
7b-v0.2-q3_K_L
3.6GB
7b-v0.2-q4_K_S
3.9GB
7b-v0.2-q4_K_M
4.1GB
7b-v0.2-q5_K_S
4.7GB
7b-v0.2-q5_K_M
4.8GB
7b-v0.2-q6_K
5.5GB
7b-v0.2-fp16
13GB
Updated 6 months ago
6 months ago
0fa68068d970 · 3.8GB
model
archllama
·
parameters7B
·
quantization4-bit
3.8GB
template
{{ .System }}
User: {{ .Prompt }}
Assistant:
45B
params
{"num_ctx":4096,"stop":["User:","Assistant:"]}
46B
Readme
Xwin-LM is a model based on Llama 2 using various techniques to improve its quality.