running-frog
commented
1 year ago 我也是同样问题，难道没有人解决吗？
Open duxiaowu opened 1 year ago
running-frog
commented
1 year ago 我也是同样问题，难道没有人解决吗？
haiwen-yang
commented
1 year ago 加1
E:\WorkSpace\IDEAProject\JittorLLMs>python clidemo.py chatglm
[i 0606 15:39:58.275000 24 compiler.py:955] Jittor(1.3.7.16) src: d:\myenvresource\python\lib\site-packages\jittor
[i 0606 15:39:58.308000 24 compiler.py:956] cl at C:\Users\BOKE.cache\jittor\msvc\VC____\\bin\cl.exe(19.29.30133)
[i 0606 15:39:58.309000 24 compiler.py:957] cache_path: C:\Users\BOKE.cache\jittor\jt1.3.7\cl\py3.10.0\Windows-10-10.x29\IntelRCoreTMi5xc1\default
[i 0606 15:39:59.266000 24 init.py:411] Found gdb(10.2) at D:\MySoftApp\TM_GccApp\bin\gdb.EXE.
[i 0606 15:39:59.375000 24 init.py:411] Found addr2line(2.36.1) at D:\MySoftApp\TM_GccApp\bin\addr2line.EXE.
[i 0606 15:39:59.405000 24 init.py:227] Total mem: 15.78GB, using 5 procs for compiling.
[i 0606 15:40:00.727000 24 jit_compiler.cc:28] Load ccpath: C:\Users\BOKE.cache\jittor\msvc\VC____\\bin\cl.exe
[i 0606 15:40:00.731000 24 py_var_tracer.cc:22] Load trace_py_var: 3
Explicitly passing a revision is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision
.
Explicitly passing a revision is encouraged when loading a configuration with custom code to ensure no malicious code has been contributed in a newer
revision.
Explicitly passing a revision is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision
.
Loading checkpoint shards: 0%| | 0/8 [00:00<?, ?it/s][
e 0606 15:40:16.682000 24 init.py:1597] load parameter transformer.word_embeddings.weight failed: expect the shape of transformer.word_embeddings.we
ight to be [150528,4096,], but got [130528,4096,]
[w 0606 15:40:16.683000 24 init.py:1599] load total 25 params, 1 failed
[e 0606 15:40:16.684000 24 init.py:1597] load parameter word_embeddings.weight failed: expect the shape of word_embeddings.weight to be [150528,4096
,], but got [130528,4096,]
[w 0606 15:40:16.684000 24 init.py:1599] load total 25 params, 1 failed
[e 0606 15:40:16.685000 24 init.py:1597] load parameter weight failed: expect the shape of weight to be [150528,4096,], but got [130528,4096,]
[w 0606 15:40:16.685000 24 init.py:1599] load total 1 params, 1 failed
Loading checkpoint shards: 88%|█████████████████████████████████████████████████████████████████████████████▉ | 7/8 [00:32<00:04, 4.91s/it][
e 0606 15:41:08.671000 24 init.py:1597] load parameter lm_head.weight failed: expect the shape of lm_head.weight to be [150528,4096,], but got [1305
28,4096,]
[w 0606 15:41:08.897000 24 init.py:1599] load total 1 params, 1 failed
[e 0606 15:41:08.914000 24 init.py:1597] load parameter weight failed: expect the shape of weight to be [150528,4096,], but got [130528,4096,]
[w 0606 15:41:08.915000 24 init.py:1599] load total 1 params, 1 failed
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████████████████████████████████████| 8/8 [00:56<00:00, 7.00s/it]
用户输入:hello
raise_cxx_exception 3221225477
[e 0606 15:55:29.260000 24 mem_info.cc:101] appear time -> node cnt: {1:678, }
Traceback (most recent call last):
File "E:\WorkSpace\IDEAProject\JittorLLMs\cli_demo.py", line 9, in
_cast__JIT_1JIT_cpu_1graph_010000_varinfohash_e8d5f58a3e108a25_op.cc [OP TYPE]: fused_op:( unary.cast,)
E:\WorkSpace\IDEAProject\JittorLLMs\cli_demo.py:8 <<module>>
E:\WorkSpace\IDEAProject\JittorLLMs\models\__init__.py:46 <get_model>
E:\WorkSpace\IDEAProject\JittorLLMs\models\chatglm\__init__.py:48 <get_model>
E:\WorkSpace\IDEAProject\JittorLLMs\models\chatglm\__init__.py:26 <__init__>
D:\MyEnvResource\Python\lib\site-packages\jittor\__init__.py:1730 <float32> [Reason]: exception: access violation writing 0000000000000000
GreatAsia
commented
1 year ago [e 0702 10:05:07.247670 00 mem_info.cc:101] appear time → node cnt: {2:494, } Traceback (most recent call last): File “/Users/zhou/Documents/JittorLLMs/JittorLLMs/cli_demo.py”, line 9, in model.chat() File “/Users/zhou/Documents/JittorLLMs/JittorLLMs/models/chatglm/init.py”, line 36, in chat for response, history in self.model.stream_chat(self.tokenizer, text, history=history): File “/Users/zhou/.cache/huggingface/modules/transformers_modules/local/modeling_chatglm.py”, line 1259, in stream_chat for outputs in self.stream_generate(input_ids, gen_kwargs): File “/Users/zhou/.cache/huggingface/modules/transformers_modules/local/modeling_chatglm.py”, line 1334, in stream_generate model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs) File “/Users/zhou/.cache/huggingface/modules/transformers_modules/local/modeling_chatglm.py”, line 1083, in prepare_inputs_for_generation mask_token = MASK if MASK in input_ids else gMASK File “/usr/local/lib/python3.9/site-packages/jittor/misc.py”, line 132, in contains return bool((x == key).any()) File “/usr/local/lib/python3.9/site-packages/jittor/init.py”, line 2013, in to_bool return ori_bool(v.item()) RuntimeError: [f 0702 10:05:07.247871 00 executor.cc:682] Execute fused operator(197/257) failed.
[OP TYPE]: fused_op:( unary.cast,)
/Users/zhou/Documents/JittorLLMs/JittorLLMs/cli_demo.py:8 <>
/Users/zhou/Documents/JittorLLMs/JittorLLMs/models/init.py:46
nashsophie
commented
1 year ago 遇到同样的问题，内存64G，显卡2060，6g显存
[e 0516 13:46:47.609943 16 mem_info.cc:101] appear time -> node cnt: {1:710, } Traceback (most recent call last): File "/home/duweibin/JittorLLMs/cli_demo.py", line 9, in
model.chat()
File "/home/duweibin/JittorLLMs/models/chatglm/init.py", line 36, in chat
for response, history in self.model.stream_chat(self.tokenizer, text, history=history):
File "/home/duweibin/.cache/huggingface/modules/transformers_modules/local/modeling_chatglm.py", line 1259, in stream_chat
for outputs in self.stream_generate(input_ids, gen_kwargs):
File "/home/duweibin/.cache/huggingface/modules/transformers_modules/local/modeling_chatglm.py", line 1334, in stream_generate
model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs)
File "/home/duweibin/.cache/huggingface/modules/transformers_modules/local/modeling_chatglm.py", line 1083, in prepare_inputs_for_generation
mask_token = MASK if MASK in input_ids else gMASK
File "/home/duweibin/.local/lib/python3.9/site-packages/jittor/misc.py", line 132, in contains
return bool((x == key).any())
File "/home/duweibin/.local/lib/python3.9/site-packages/jittor/init.py", line 2013, in to_bool
return ori_bool(v.item())
RuntimeError: [f 0516 13:46:47.610098 16 executor.cc:682]
Execute fused operator(89/257) failed.
[OP TYPE]: fused_op:( unary.cast,)
[Reason]: [f 0516 13:46:47.012486 16 swap.cc:63] Check failed: res==1 failed to write swap file /home/duweibin/.cache/jittor/jt1.3.7/g++7.5.0/py3.9.16/Linux-5.4.0-13xf6/IntelRCoreTMi5x26/default/tmp/29197-11719.bin 0 268435456 Var(11719:1:1:1:i0:o0:s1:n0,float32,transformer.layers.9.mlp.dense_h_to_4h.weight,7f3fb1ffb040)[16384,4096,]{/home/duweibin/JittorLLMs/cli_demo.py:8() -> /home/duweibin/JittorLLMs/models/init.py:46(get_model) -> /home/duweibin/JittorLLMs/models/chatglm/init.py:48(get_model) -> /home/duweibin/JittorLLMs/models/chatglm/init.py:26(init) -> /home/duweibin/.local/lib/python3.9/site-packages/jittor/init.py:1730(float32) -> }