wli1995 commited on
Commit
911f08e
·
verified ·
1 Parent(s): f61ff15

update axmodel and demo

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +30 -0
  2. README.md +43 -51
  3. deepseek-r1-1.5b-ax650/qwen2_p128_l0_together.axmodel +2 -2
  4. deepseek-r1-1.5b-ax650/qwen2_p128_l10_together.axmodel +2 -2
  5. deepseek-r1-1.5b-ax650/qwen2_p128_l11_together.axmodel +2 -2
  6. deepseek-r1-1.5b-ax650/qwen2_p128_l12_together.axmodel +2 -2
  7. deepseek-r1-1.5b-ax650/qwen2_p128_l13_together.axmodel +2 -2
  8. deepseek-r1-1.5b-ax650/qwen2_p128_l14_together.axmodel +2 -2
  9. deepseek-r1-1.5b-ax650/qwen2_p128_l15_together.axmodel +2 -2
  10. deepseek-r1-1.5b-ax650/qwen2_p128_l16_together.axmodel +2 -2
  11. deepseek-r1-1.5b-ax650/qwen2_p128_l17_together.axmodel +2 -2
  12. deepseek-r1-1.5b-ax650/qwen2_p128_l18_together.axmodel +2 -2
  13. deepseek-r1-1.5b-ax650/qwen2_p128_l19_together.axmodel +2 -2
  14. deepseek-r1-1.5b-ax650/qwen2_p128_l1_together.axmodel +2 -2
  15. deepseek-r1-1.5b-ax650/qwen2_p128_l20_together.axmodel +2 -2
  16. deepseek-r1-1.5b-ax650/qwen2_p128_l21_together.axmodel +2 -2
  17. deepseek-r1-1.5b-ax650/qwen2_p128_l22_together.axmodel +2 -2
  18. deepseek-r1-1.5b-ax650/qwen2_p128_l23_together.axmodel +2 -2
  19. deepseek-r1-1.5b-ax650/qwen2_p128_l24_together.axmodel +2 -2
  20. deepseek-r1-1.5b-ax650/qwen2_p128_l25_together.axmodel +2 -2
  21. deepseek-r1-1.5b-ax650/qwen2_p128_l26_together.axmodel +2 -2
  22. deepseek-r1-1.5b-ax650/qwen2_p128_l27_together.axmodel +2 -2
  23. deepseek-r1-1.5b-ax650/qwen2_p128_l2_together.axmodel +2 -2
  24. deepseek-r1-1.5b-ax650/qwen2_p128_l3_together.axmodel +2 -2
  25. deepseek-r1-1.5b-ax650/qwen2_p128_l4_together.axmodel +2 -2
  26. deepseek-r1-1.5b-ax650/qwen2_p128_l5_together.axmodel +2 -2
  27. deepseek-r1-1.5b-ax650/qwen2_p128_l6_together.axmodel +2 -2
  28. deepseek-r1-1.5b-ax650/qwen2_p128_l7_together.axmodel +2 -2
  29. deepseek-r1-1.5b-ax650/qwen2_p128_l8_together.axmodel +2 -2
  30. deepseek-r1-1.5b-ax650/qwen2_p128_l9_together.axmodel +2 -2
  31. deepseek-r1-1.5b-ax650/qwen2_post.axmodel +1 -1
  32. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l0_together.axmodel +3 -0
  33. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l10_together.axmodel +3 -0
  34. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l11_together.axmodel +3 -0
  35. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l12_together.axmodel +3 -0
  36. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l13_together.axmodel +3 -0
  37. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l14_together.axmodel +3 -0
  38. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l15_together.axmodel +3 -0
  39. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l16_together.axmodel +3 -0
  40. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l17_together.axmodel +3 -0
  41. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l18_together.axmodel +3 -0
  42. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l19_together.axmodel +3 -0
  43. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l1_together.axmodel +3 -0
  44. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l20_together.axmodel +3 -0
  45. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l21_together.axmodel +3 -0
  46. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l22_together.axmodel +3 -0
  47. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l23_together.axmodel +3 -0
  48. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l24_together.axmodel +3 -0
  49. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l25_together.axmodel +3 -0
  50. deepseek-r1-1.5b-int4-ax650/qwen2_p128_l26_together.axmodel +3 -0
.gitattributes CHANGED
@@ -97,3 +97,33 @@ figures/figures_benchmark.jpg filter=lfs diff=lfs merge=lfs -text
97
  figures/benchmark.jpg filter=lfs diff=lfs merge=lfs -text
98
  main_axcl_aarch64 filter=lfs diff=lfs merge=lfs -text
99
  main_axcl_x86 filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
97
  figures/benchmark.jpg filter=lfs diff=lfs merge=lfs -text
98
  main_axcl_aarch64 filter=lfs diff=lfs merge=lfs -text
99
  main_axcl_x86 filter=lfs diff=lfs merge=lfs -text
100
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l0_together.axmodel filter=lfs diff=lfs merge=lfs -text
101
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l10_together.axmodel filter=lfs diff=lfs merge=lfs -text
102
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l11_together.axmodel filter=lfs diff=lfs merge=lfs -text
103
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l12_together.axmodel filter=lfs diff=lfs merge=lfs -text
104
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l13_together.axmodel filter=lfs diff=lfs merge=lfs -text
105
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l14_together.axmodel filter=lfs diff=lfs merge=lfs -text
106
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l15_together.axmodel filter=lfs diff=lfs merge=lfs -text
107
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l16_together.axmodel filter=lfs diff=lfs merge=lfs -text
108
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l17_together.axmodel filter=lfs diff=lfs merge=lfs -text
109
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l18_together.axmodel filter=lfs diff=lfs merge=lfs -text
110
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l19_together.axmodel filter=lfs diff=lfs merge=lfs -text
111
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l1_together.axmodel filter=lfs diff=lfs merge=lfs -text
112
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l20_together.axmodel filter=lfs diff=lfs merge=lfs -text
113
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l21_together.axmodel filter=lfs diff=lfs merge=lfs -text
114
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l22_together.axmodel filter=lfs diff=lfs merge=lfs -text
115
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l23_together.axmodel filter=lfs diff=lfs merge=lfs -text
116
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l24_together.axmodel filter=lfs diff=lfs merge=lfs -text
117
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l25_together.axmodel filter=lfs diff=lfs merge=lfs -text
118
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l26_together.axmodel filter=lfs diff=lfs merge=lfs -text
119
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l27_together.axmodel filter=lfs diff=lfs merge=lfs -text
120
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l2_together.axmodel filter=lfs diff=lfs merge=lfs -text
121
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l3_together.axmodel filter=lfs diff=lfs merge=lfs -text
122
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l4_together.axmodel filter=lfs diff=lfs merge=lfs -text
123
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l5_together.axmodel filter=lfs diff=lfs merge=lfs -text
124
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l6_together.axmodel filter=lfs diff=lfs merge=lfs -text
125
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l7_together.axmodel filter=lfs diff=lfs merge=lfs -text
126
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l8_together.axmodel filter=lfs diff=lfs merge=lfs -text
127
+ deepseek-r1-1.5b-int4-ax650/qwen2_p128_l9_together.axmodel filter=lfs diff=lfs merge=lfs -text
128
+ deepseek-r1-1.5b-int4-ax650/qwen2_post.axmodel filter=lfs diff=lfs merge=lfs -text
129
+ main_ax650 filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -18,7 +18,7 @@ This version of DeepSeek-R1-Distill-Qwen-1.5B has been converted to run on the A
18
 
19
  This model has been optimized with the following LoRA:
20
 
21
- Compatible with Pulsar2 version: 4.1
22
 
23
  ## Feature
24
 
@@ -28,7 +28,7 @@ Compatible with Pulsar2 version: 4.1
28
 
29
  ## Convert tools links:
30
 
31
- For those who are interested in model conversion, you can try to export axmodel through the original repo : https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
32
 
33
  [Pulsar2 Link, How to Convert LLM from Huggingface to axmodel](https://pulsar2-docs.readthedocs.io/en/latest/appendix/build_llm.html)
34
 
@@ -79,26 +79,32 @@ Download all files from this repository to the device
79
  ```
80
  root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# tree -L 1
81
  .
 
82
  |-- config.json
83
- |-- deepseek-r1-1.5b-ctx-ax650
 
84
  |-- deepseek-r1_tokenizer
85
- |-- deepseek-r1_tokenizer_uid.py
 
86
  |-- main_ax650
87
  |-- main_axcl_aarch64
88
  |-- main_axcl_x86
89
  |-- post_config.json
90
- |-- run_deepseek-r1_1.5b_ctx_ax650.sh
91
- |-- run_deepseek-r1_1.5b_ctx_axcl_aarch64.sh
92
- `-- run_deepseek-r1_1.5b_ctx_axcl_x86.sh
93
-
94
- 2 directories, 9 files
 
 
 
95
  ```
96
 
97
  #### Start the Tokenizer service
98
 
99
  ```
100
  root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# python3 deepseek-r1_tokenizer_uid.py
101
- Server running at http://0.0.0.0:12345
102
  ```
103
 
104
  #### System prompt cache
@@ -108,14 +114,13 @@ Server running at http://0.0.0.0:12345
108
  - This folder needs to be created manually before running, for example `mkdir kvcache`
109
 
110
  ```
111
- root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# cat run_deepseek-r1_1.5b_ctx_ax650.sh
112
  ./main_ax650 \
113
- --template_filename_axmodel "deepseek-r1-1.5b-ctx-ax650/qwen2_p128_l%d_together.axmodel" \
114
  --axmodel_num 28 \
115
- --tokenizer_type 2 \
116
- --url_tokenizer_model "http://0.0.0.0:12345" \
117
- --filename_post_axmodel "deepseek-r1-1.5b-ctx-ax650/qwen2_post.axmodel" \
118
- --filename_tokens_embed "deepseek-r1-1.5b-ctx-ax650/model.embed_tokens.weight.bfloat16.bin" \
119
  --tokens_embed_num 151936 \
120
  --tokens_embed_size 1536 \
121
  --use_mmap_load_embed 1 \
@@ -124,18 +129,16 @@ root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# cat run_deepseek-r1_1.5b_ct
124
 
125
  #### Inference with AX650 Host, such as M4N-Dock(爱芯派Pro) or AX650N DEMO Board
126
 
127
- Open another terminal and run `run_deepseek-r1_1.5b_ctx_ax650.sh`
128
 
129
  ```
130
- root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# ./run_deepseek-r1_1.5b_ctx_ax650.sh
131
  [I][ Init][ 110]: LLM init start
132
- [I][ Init][ 34]: connect http://0.0.0.0:12345 ok
133
- [I][ Init][ 57]: uid: 7fedc3e5-e824-4915-935a-c0de5a341928
134
  bos_id: 151646, eos_id: 151643
135
- 3% | ██ | 1 / 31 [2.28s<70.62s, 0.44 count/s] tokenizer init ok
136
- [I][ Init][ 26]: LLaMaEmbedSelector use mmap
137
- 100% | ████████████████████████████████ | 31 / 31 [26.47s<26.47s, 1.17 count/s] init post axmodel ok,remain_cmm(8947 MB)
138
- [I][ Init][ 188]: max_token_len : 2047
139
  [I][ Init][ 193]: kv_cache_size : 256, kv_cache_num: 2047
140
  [I][ Init][ 201]: prefill_token_num : 128
141
  [I][ Init][ 205]: grp: 1, prefill_max_token_num : 1
@@ -147,11 +150,7 @@ bos_id: 151646, eos_id: 151643
147
  [I][ Init][ 205]: grp: 7, prefill_max_token_num : 768
148
  [I][ Init][ 205]: grp: 8, prefill_max_token_num : 896
149
  [I][ Init][ 205]: grp: 9, prefill_max_token_num : 1024
150
- [I][ Init][ 205]: grp: 10, prefill_max_token_num : 1152
151
- [I][ Init][ 205]: grp: 11, prefill_max_token_num : 1280
152
- [I][ Init][ 205]: grp: 12, prefill_max_token_num : 1408
153
- [I][ Init][ 205]: grp: 13, prefill_max_token_num : 1536
154
- [I][ Init][ 209]: prefill_max_token_num : 1536
155
  [I][ load_config][ 282]: load config:
156
  {
157
  "enable_repetition_penalty": false,
@@ -167,31 +166,24 @@ bos_id: 151646, eos_id: 151643
167
 
168
  [I][ Init][ 218]: LLM init ok
169
  Type "q" to exit, Ctrl+c to stop current running
170
- [I][ GenerateKVCachePrefill][ 271]: input token num : 16, prefill_split_num : 1 prefill_grpid : 2
171
- [I][ GenerateKVCachePrefill][ 308]: input_num_token:16
172
- [I][ main][ 230]: precompute_len: 16
173
- [I][ main][ 231]: system_prompt:
174
- prompt >> 1+2=?
175
- [I][ SetKVCache][ 531]: prefill_grpid:2 kv_cache_num:128 precompute_len:16 input_num_token:8
176
- [I][ SetKVCache][ 534]: current prefill_max_token_num:1408
177
- [I][ Run][ 660]: input token num : 8, prefill_split_num : 1
178
- [I][ Run][ 686]: input_num_token:8
179
- [I][ Run][ 829]: ttft: 306.60 ms
180
  <think>
181
- Okay, the user has asked "1+2=?", which is a simple addition question.
182
- I should provide the answer, but also consider if there's more to it.
183
-
184
- Since the user specified "Qwen, created by Alibaba Cloud,"
185
- maybe they're testing if I understand the context or need further assistance within that framework.
186
-
187
- I'll give the correct sum and let them know if they need anything else. That should be helpful.
188
  </think>
189
 
190
- 1 + 2 equals **3**.
 
 
191
 
192
- [N][ Run][ 943]: hit eos,avg 11.25 token/s
193
 
194
- [I][ GetKVCache][ 500]: precompute_len:123, remaining:1413
195
- prompt >> q
196
- root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx#
197
  ```
 
18
 
19
  This model has been optimized with the following LoRA:
20
 
21
+ Compatible with Pulsar2 version: 4.2
22
 
23
  ## Feature
24
 
 
28
 
29
  ## Convert tools links:
30
 
31
+ For those who are interested in model conversion, you can try to export axmodel through the original repo : https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B and https://huggingface.co/jakiAJK/DeepSeek-R1-Distill-Qwen-1.5B_GPTQ-int4
32
 
33
  [Pulsar2 Link, How to Convert LLM from Huggingface to axmodel](https://pulsar2-docs.readthedocs.io/en/latest/appendix/build_llm.html)
34
 
 
79
  ```
80
  root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# tree -L 1
81
  .
82
+ |-- README.md
83
  |-- config.json
84
+ |-- deepseek-r1-1.5b-ax650
85
+ |-- deepseek-r1-1.5b-int4-ax650
86
  |-- deepseek-r1_tokenizer
87
+ |-- deepseek-r1_tokenizer.py
88
+ |-- figures
89
  |-- main_ax650
90
  |-- main_axcl_aarch64
91
  |-- main_axcl_x86
92
  |-- post_config.json
93
+ |-- run_deepseek-r1_1.5B_ax650.sh
94
+ |-- run_deepseek-r1_1.5B_int4_ax650.sh
95
+ |-- run_deepseek-r1_1.5b_axcl_aarch64.sh
96
+ |-- run_deepseek-r1_1.5b_axcl_x86.sh
97
+ |-- run_deepseek-r1_1.5b_int4_axcl_aarch64.sh
98
+ `-- run_deepseek-r1_1.5b_int4_axcl_x86.sh
99
+
100
+ 4 directories, 16 files
101
  ```
102
 
103
  #### Start the Tokenizer service
104
 
105
  ```
106
  root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# python3 deepseek-r1_tokenizer_uid.py
107
+ Server running at http://127.0.0.1:12345
108
  ```
109
 
110
  #### System prompt cache
 
114
  - This folder needs to be created manually before running, for example `mkdir kvcache`
115
 
116
  ```
117
+ root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# cat run_deepseek-r1_1.5b_ax650.sh
118
  ./main_ax650 \
119
+ --template_filename_axmodel "deepseek-r1-1.5b-ax650/qwen2_p128_l%d_together.axmodel" \
120
  --axmodel_num 28 \
121
+ --url_tokenizer_model "http://127.0.0.1:12345" \
122
+ --filename_post_axmodel "deepseek-r1-1.5b-ax650/qwen2_post.axmodel" \
123
+ --filename_tokens_embed "deepseek-r1-1.5b-ax650/model.embed_tokens.weight.bfloat16.bin" \
 
124
  --tokens_embed_num 151936 \
125
  --tokens_embed_size 1536 \
126
  --use_mmap_load_embed 1 \
 
129
 
130
  #### Inference with AX650 Host, such as M4N-Dock(爱芯派Pro) or AX650N DEMO Board
131
 
132
+ Open another terminal and run `run_deepseek-r1_1.5b_ax650.sh`
133
 
134
  ```
135
+ root@ax650:/mnt/qtang/llm-test/deepseek-r1-1.5b-ctx# ./run_deepseek-r1_1.5b_ax650.sh
136
  [I][ Init][ 110]: LLM init start
137
+ [I][ Init][ 34]: connect http://127.0.0.1:12345 ok
138
+ [I][ Init][ 57]: uid: 96cbe293-ff24-4011-b3a1-b5ab95234990
139
  bos_id: 151646, eos_id: 151643
140
+ 3% | ██ | 1 / 31 [2.07s<64.23s, 0.48 count/s] tokenizer init ok[I][ Init][ 26]: LLaMaEmbedSelector use mmap
141
+ 100% | ████████████████████████████████ | 31 / 31 [8.76s<8.76s, 3.54 count/s] init post axmodel ok,remain_cmm(8114 MB)[I][ Init][ 188]: max_token_len : 2047
 
 
142
  [I][ Init][ 193]: kv_cache_size : 256, kv_cache_num: 2047
143
  [I][ Init][ 201]: prefill_token_num : 128
144
  [I][ Init][ 205]: grp: 1, prefill_max_token_num : 1
 
150
  [I][ Init][ 205]: grp: 7, prefill_max_token_num : 768
151
  [I][ Init][ 205]: grp: 8, prefill_max_token_num : 896
152
  [I][ Init][ 205]: grp: 9, prefill_max_token_num : 1024
153
+ [I][ Init][ 209]: prefill_max_token_num : 1024
 
 
 
 
154
  [I][ load_config][ 282]: load config:
155
  {
156
  "enable_repetition_penalty": false,
 
166
 
167
  [I][ Init][ 218]: LLM init ok
168
  Type "q" to exit, Ctrl+c to stop current running
169
+ [I][ GenerateKVCachePrefill][ 275]: input token num : 11, prefill_split_num : 1 prefill_grpid : 2
170
+ [I][ GenerateKVCachePrefill][ 315]: input_num_token:11
171
+ [I][ main][ 228]: precompute_len: 11
172
+ [I][ main][ 229]: system_prompt:
173
+ prompt >> 你是谁
174
+ [I][ SetKVCache][ 529]: prefill_grpid:2 kv_cache_num:128 precompute_len:11 input_num_token:6
175
+ [I][ SetKVCache][ 532]: current prefill_max_token_num:896
176
+ [I][ Run][ 658]: input token num : 6, prefill_split_num : 1
177
+ [I][ Run][ 684]: input_num_token:6
178
+ [I][ Run][ 807]: ttft: 256.87 ms
179
  <think>
180
+ 您好!我是由中国的深度求索(DeepSeek)公司开发的智能助手DeepSeek-R1。如您有任何任何问题,我会尽我所能为您提供帮助。
 
 
 
 
 
 
181
  </think>
182
 
183
+ 您好!我是由中国的深度求索(DeepSeek)公司开发的智能助手DeepSeek-R1。如您有任何任何问题,我会尽我所能为您提供帮助。
184
+
185
+ [N][ Run][ 921]: hit eos,avg 17.68 token/s
186
 
187
+ [I][ GetKVCache][ 498]: precompute_len:91, remaining:933
188
 
 
 
 
189
  ```
deepseek-r1-1.5b-ax650/qwen2_p128_l0_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:522ade59d65cff80d29c70a7dfc6a5dc033d603974d74dcd19b354b09530722d
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be25a034f27581cd0f7af207a9c0598381e1f19ba0ecfdab943f56c52b1afc9a
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l10_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:459fa50ffb11858a43b54ce215476589fc4faf8dfb96ca9214b8293b58ac1103
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f1c86026e1bc725957c6e933da8760062715047b794ace7e2c1f1e7b62b16c12
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l11_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a58561902aed842a82fe0f7b2272d213a91c275ae76eca221590d08d2ff97661
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bce863f600e20101c5827dd217f0675ef9ecdb91c2b1daa6df79a3144ea4654d
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l12_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0cf81bd3b3ac80888246bfc3bc5bb297acbebd8a507ebfb791e53f4600c30ef3
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79b13a28b5d799c27df905e5165d0acfa66e9cae166bc5a0770f111b741efd79
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l13_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:986e45505669199a5715d64086defa9268ab7c1ee284ee515d65458fe9b62594
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e99ab77861a2d0db96009c64d0e369c2333173044b97affdcc626c9d70fab1a8
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l14_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:70e051d1037248b958467d58629514e627d964b2f27ec6d6c9cbce72fd1b152a
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa2518bf3b7f068dbe709a3cfea3907d49096ce9fae393673957e88adbbdb913
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l15_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5760efccff8171cff54e9c7b299558cd437b9e8ca2d27ad67ebf7ce7cfae46d6
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a71aa34f0cb5c912058a57e051d0f06ef2baad10f47fc21700c96bdeae011de
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l16_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f12844f3420f944f332ca51033b912ee4de89be6c7ded26c58e44dc24d2724e9
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b14d934b5ed86b85b403ad5b7f920e5d7d7327c5d9d9625e453dd54f20b96057
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l17_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:53a18582cd841cd56d8039e739feafeae2051d84d1f036b5b6ac5791186c47a2
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c76091b5259d4c46c31f083806edba162c268af509ee84a4aebfcb4c29686090
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l18_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:283a07ecb1670ad7198a42260d3444696ae2e16a0eb4ce9e600678c1aab63657
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa0b4f01e430bfb6b2b9d147b9b987028f9abeddba6afe92c330d1fad89785d7
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l19_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:09fa3121fbcbf0c003cc8c2731030b1e56af41defd9068824d0f2be128eea9ba
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e240dae492739f1911c7ff5ab7190454684943cce1044c3566cac711cafbd5e0
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l1_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f2327972616360b4b7bc222f5fd17de676def7bf049c49b939f4057cb37591fd
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a08ea7b97e094625a8be0a08d2be0d8e54960f68cadde74adbc3f8039d375d3
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l20_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:96ff64967a922794b9ef77df7b23d515cf7b892805c1ce4fa4abe84be6cdd971
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6fdb4f7da5a7514a260e4f497f5db4af0ce94d5b6ad229fb49a8b808be4bce6b
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l21_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ea512f881311fdeb068f02138cf4efa1784ae2d0f685c31c7290cb464a9f6ed8
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a52db67d8d790ceaf9c44604d3206d1207955d92b288f319c009c605fbff9e6
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l22_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:495fccb3b187ac3b99cdcd1b7b50905ac4b820b3c9e20358ae2fceaac399284c
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06b8c833345060de714dddbec563abaf98194537e0d4932a47feee727dc17939
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l23_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3e29d32050388048d926a1d7e2bb32b3b3f30778db9cabcdbc1fc732e9c10bcb
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a747e0072af4c386af7b8c4f46388732572a8bab4b72cb3df9efb4dee52054db
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l24_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f0641564f5017fdaf98bbbfff5364b88e0911c5d3a343585ef95ffbdf4d76628
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5591329fbdb9e2e0c939b44ae09796048b621bc7910fcb41eb2e96f1aeea5848
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l25_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:38bbfc9f023ba81ad8e778c179208760a091b02911a54e749ff2b24761bb0bec
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f3f13627396c2e87f4bbfff0c449f7d51f45ece7878cad12fe185a580d9a83e
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l26_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:97abb5dd5d4800be31950475c9769a6025382e834bebf9eebd8ec6b86159c550
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54119e40e581a543fd4144afff911ae4a4da0990b31c796737f6bed72dc9fa10
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l27_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:762fce5404016b3f3aff3a0d66f587f0ac06d4176c0e1ddeca68885eca0029ff
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20f2701ef879f3690ee7a871097f0388224dbec76c2af2c0331ec13d7cef5a58
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l2_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:66d0843374ec2aa65842474f5146b90339a962541af9c54d60a0909f6107ef0a
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c69bb6cc74eec834aeff2bb9cb5d1deae54f3c805b345cc16d78697d6a745b2
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l3_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ca361332c875422394d3e11b59c0c8069b9f66692fe6a80533021f2f0444b98f
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d66e85cb4980ba39bc4205d04d30a6030f3e07f20a2118dd3edbac5b800eceee
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l4_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:00cf57979c4ccc3eb6e9e7d1b701dd44fde3e64f528a19d8924e390aafd43b97
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb2e65b19a8474ea35aa24bd1ab417df4890cceedb08ae5ea33d4bef2d537f9f
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l5_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0e34467ca89cdbd0dc15dfb83be967eb4bcc217604d72a82793ce1ef011e84be
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:443f4dfe91116c071a37f2d92464bd953e56c08081c9f8b4f89498f591962100
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l6_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:df34fdfd121d439ecc3f4c2f1965004bac72b3bf1e467a4fc10023a6bea8d5db
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7dca045308029638415dc8b816b9b5a256d64d21384c4b71b0ad058dae450420
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l7_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2858db26c2692db1339be9a09e8c95a59e9ae66033a9c397f2ba84cd4aac864e
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da0519e123bc0f7ab4680f7120e462a08f2540768edb82f904cfc8a0bc25b580
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l8_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:da4c317a23fce0ae9c6353427e00f33db5d9866ed1ec4d389170b1d0eb88ba9f
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3dd7537c5436cda4f3773ecf274348995166b1e1de9af26180dfec708f35c71
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_p128_l9_together.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ec1ed959ab0525f9bc2f40deee81b6169296a11acce9472c1830792b15b101a4
3
- size 52834948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9795532614c166985165869e63023c82a96f61912980f3a304a96f49b781a55
3
+ size 62970116
deepseek-r1-1.5b-ax650/qwen2_post.axmodel CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ede68b189b7892beca2a4877bf36c9a7f98ee420c93fa46a5e106fc545b78593
3
  size 254692887
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:706dcdd770eda42a82548f3d3c5cc9dd3840818d72eca312bd96ed7ac9e6714d
3
  size 254692887
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l0_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6aaa5ceb34564e043555a0b783be09c637ba6f7474e699c089e0b66ee1d25439
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l10_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85a6f545503d31dca0214bad55439f0c50399b7258c70d2e274b8d414cab8ff8
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l11_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3be0a3d96470fa7682862373f90320a673dfeb1e7505c4640d62070b6ee20dd7
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l12_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:260b2593262480cbd1515f60ac45d7c3d350e80a9a3529606f49106baf662d33
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l13_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a75ee8dcf1bb19f28592f2c11a2c8b0c5bc7cedef3124182c9778e313034a6e
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l14_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3375832c9e6b5c94185353906b5de4c88e26fc966e904090aa12029560b7509
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l15_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5482b12cee22649f45e9d19a9f3436862c0175f71bcaecd45b5de95029b609c
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l16_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4061bbde2821e282d15ea5a23e95a9ecc6c463f2bfe2592845f8ac4d96d62ec6
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l17_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f18b66ae4a5bafac5bde713de4cbcdf7a6a0a46cae9b7c7ac23e259cd741caa1
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l18_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea7c7663e96ab4062c53c21f3b2f9abaf7fd3527b53be23759272155836ce654
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l19_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f201ffe7f99797eccc9bec0725c7ca83ff7e6c541d9d84eaac117681e4de1d77
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l1_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:16641c20d5633676bf80bae34a6c66763d8f0b23f708ff83ace07f2a04fc55df
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l20_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdb618b59cf9d8d63a5cfbdaec59c72c7be233d28941e42e2536dd9eb2ad08f9
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l21_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:318e348c4cd4be6825e11e263c34653fbdefa61b7b8a80ef1fae18872ec2b7f4
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l22_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dfcecc0aefe64288f6e70a1257b7ce97240d3415dccd1a0e784412a64aa996bd
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l23_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18b79890f3392302c71d40d90a853362bf97b7ebec020836e1916c71dbc6b73e
3
+ size 37866696
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l24_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b887c92a69a95207c14ace97a2540c14d6ac3d9be598accddb121bc2d81769a
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l25_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30d46409a6dcb7f47271f1308e1698346ab6658d6cc41ac28dee36863fe0ee2a
3
+ size 37866688
deepseek-r1-1.5b-int4-ax650/qwen2_p128_l26_together.axmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a8c666022bcb30590dd4d61a77ab688faa085d3bcb7ead81f7ba628161f697f
3
+ size 37866688