Unsloth GRPO datta script

gistfile1.txt

Unsloth 2025.11.3 patched 36 layers with 36 QKV layers, 36 O layers and 36 MLP layers.
Performing substitution for additional_keys=set()
Unsloth: Just some info: will skip parsing ['attention_norm', 'cross_attn_input_layernorm', 'cross_attn_post_attention_layernorm', 'post_layernorm', 'norm2', 'post_attention_layernorm', 'input_layernorm', 'pre_feedforward_layernorm', 'norm', 'post_feedforward_layernorm', 'k_norm', 'layer_norm1', 'q_norm', 'layer_norm2', 'ffn_norm', 'norm1']
Unsloth: Tokenizing ["text"] (num_proc=64): 100%|██████████| 1126/1126 [00:11<00:00, 95.40 examples/s] 
==((====))==  Unsloth - 2x faster free finetuning | Num GPUs used = 1
   \\   /|    Num examples = 1,126 | Num Epochs = 1 | Total steps = 100
O^O/ \_/ \    Batch size per device = 4 | Gradient accumulation steps = 1
\        /    Data Parallel GPUs = 1 | Total batch size (4 x 1 x 1) = 4
 "-____-"     Trainable parameters = 87,293,952 of 8,278,029,312 (1.05% trained)
{'loss': 0.5516, 'grad_norm': 0.5312474966049194, 'learning_rate': 0.0, 'epoch': 0.0}
Unsloth: Will smartly offload gradients to save VRAM!
{'loss': 0.5751, 'grad_norm': 0.4997809827327728, 'learning_rate': 4e-05, 'epoch': 0.01}
{'loss': 0.6654, 'grad_norm': 0.5570040941238403, 'learning_rate': 8e-05, 'epoch': 0.01}
{'loss': 0.621, 'grad_norm': 0.49220409989356995, 'learning_rate': 0.00012, 'epoch': 0.01}
{'loss': 0.4791, 'grad_norm': 0.3357393741607666, 'learning_rate': 0.00016, 'epoch': 0.02}
{'loss': 0.5139, 'grad_norm': 0.23720338940620422, 'learning_rate': 0.0002, 'epoch': 0.02}
{'loss': 0.4249, 'grad_norm': 0.22417885065078735, 'learning_rate': 0.00019789473684210526, 'epoch': 0.02}
{'loss': 0.4403, 'grad_norm': 0.22005413472652435, 'learning_rate': 0.00019578947368421054, 'epoch': 0.03}
{'loss': 0.3843, 'grad_norm': 0.2089555263519287, 'learning_rate': 0.0001936842105263158, 'epoch': 0.03}
{'loss': 0.3661, 'grad_norm': 0.1853521764278412, 'learning_rate': 0.00019157894736842104, 'epoch': 0.04}
{'loss': 0.3368, 'grad_norm': 0.23009157180786133, 'learning_rate': 0.00018947368421052632, 'epoch': 0.04}
{'loss': 0.344, 'grad_norm': 0.19669091701507568, 'learning_rate': 0.0001873684210526316, 'epoch': 0.04}
{'loss': 0.3621, 'grad_norm': 0.18337242305278778, 'learning_rate': 0.00018526315789473685, 'epoch': 0.05}
{'loss': 0.3082, 'grad_norm': 0.15389038622379303, 'learning_rate': 0.0001831578947368421, 'epoch': 0.05}
{'loss': 0.4513, 'grad_norm': 0.19309887290000916, 'learning_rate': 0.00018105263157894739, 'epoch': 0.05}
{'loss': 0.253, 'grad_norm': 0.1269475668668747, 'learning_rate': 0.00017894736842105264, 'epoch': 0.06}
{'loss': 0.3247, 'grad_norm': 0.14577877521514893, 'learning_rate': 0.0001768421052631579, 'epoch': 0.06}
{'loss': 0.2953, 'grad_norm': 0.15310949087142944, 'learning_rate': 0.00017473684210526317, 'epoch': 0.06}
{'loss': 0.3271, 'grad_norm': 0.12097445875406265, 'learning_rate': 0.00017263157894736842, 'epoch': 0.07}
{'loss': 0.3143, 'grad_norm': 0.17251430451869965, 'learning_rate': 0.0001705263157894737, 'epoch': 0.07}
{'loss': 0.3824, 'grad_norm': 0.16023662686347961, 'learning_rate': 0.00016842105263157895, 'epoch': 0.07}
{'loss': 0.3827, 'grad_norm': 0.12935973703861237, 'learning_rate': 0.00016631578947368423, 'epoch': 0.08}
{'loss': 0.3651, 'grad_norm': 0.14019690454006195, 'learning_rate': 0.00016421052631578948, 'epoch': 0.08}
{'loss': 0.314, 'grad_norm': 0.12898799777030945, 'learning_rate': 0.00016210526315789473, 'epoch': 0.09}
{'loss': 0.3622, 'grad_norm': 0.141513392329216, 'learning_rate': 0.00016, 'epoch': 0.09}
{'loss': 0.2899, 'grad_norm': 0.12062196433544159, 'learning_rate': 0.00015789473684210527, 'epoch': 0.09}
{'loss': 0.3448, 'grad_norm': 0.12003778666257858, 'learning_rate': 0.00015578947368421052, 'epoch': 0.1}
{'loss': 0.3252, 'grad_norm': 0.12177414447069168, 'learning_rate': 0.0001536842105263158, 'epoch': 0.1}
{'loss': 0.2917, 'grad_norm': 0.11223629862070084, 'learning_rate': 0.00015157894736842108, 'epoch': 0.1}
{'loss': 0.389, 'grad_norm': 0.12914486229419708, 'learning_rate': 0.00014947368421052633, 'epoch': 0.11}
{'loss': 0.3577, 'grad_norm': 0.12752953171730042, 'learning_rate': 0.00014736842105263158, 'epoch': 0.11}
{'loss': 0.3119, 'grad_norm': 0.12948615849018097, 'learning_rate': 0.00014526315789473686, 'epoch': 0.11}
{'loss': 0.2777, 'grad_norm': 0.11508759111166, 'learning_rate': 0.0001431578947368421, 'epoch': 0.12}
{'loss': 0.3316, 'grad_norm': 0.12426915019750595, 'learning_rate': 0.00014105263157894736, 'epoch': 0.12}
{'loss': 0.3592, 'grad_norm': 0.13112099468708038, 'learning_rate': 0.00013894736842105264, 'epoch': 0.12}
{'loss': 0.3531, 'grad_norm': 0.12065130472183228, 'learning_rate': 0.0001368421052631579, 'epoch': 0.13}
{'loss': 0.2714, 'grad_norm': 0.11886139959096909, 'learning_rate': 0.00013473684210526317, 'epoch': 0.13}
{'loss': 0.3196, 'grad_norm': 0.13051238656044006, 'learning_rate': 0.00013263157894736842, 'epoch': 0.13}
{'loss': 0.2932, 'grad_norm': 0.11254164576530457, 'learning_rate': 0.0001305263157894737, 'epoch': 0.14}
{'loss': 0.3292, 'grad_norm': 0.12476981431245804, 'learning_rate': 0.00012842105263157895, 'epoch': 0.14}
{'loss': 0.3437, 'grad_norm': 0.14819224178791046, 'learning_rate': 0.0001263157894736842, 'epoch': 0.15}
{'loss': 0.3994, 'grad_norm': 0.1342942863702774, 'learning_rate': 0.00012421052631578949, 'epoch': 0.15}
{'loss': 0.3108, 'grad_norm': 0.11615261435508728, 'learning_rate': 0.00012210526315789474, 'epoch': 0.15}
{'loss': 0.287, 'grad_norm': 0.10614672303199768, 'learning_rate': 0.00012, 'epoch': 0.16}
{'loss': 0.3777, 'grad_norm': 0.1580209881067276, 'learning_rate': 0.00011789473684210525, 'epoch': 0.16}
{'loss': 0.2382, 'grad_norm': 0.11780590564012527, 'learning_rate': 0.00011578947368421053, 'epoch': 0.16}
{'loss': 0.345, 'grad_norm': 0.12982822954654694, 'learning_rate': 0.0001136842105263158, 'epoch': 0.17}
{'loss': 0.2807, 'grad_norm': 0.11291559785604477, 'learning_rate': 0.00011157894736842105, 'epoch': 0.17}
{'loss': 0.3361, 'grad_norm': 0.1301521211862564, 'learning_rate': 0.00010947368421052633, 'epoch': 0.17}
{'loss': 0.3255, 'grad_norm': 0.12927880883216858, 'learning_rate': 0.00010736842105263158, 'epoch': 0.18}
{'loss': 0.3107, 'grad_norm': 0.12102185189723969, 'learning_rate': 0.00010526315789473685, 'epoch': 0.18}
{'loss': 0.2714, 'grad_norm': 0.1021929681301117, 'learning_rate': 0.00010315789473684211, 'epoch': 0.18}
{'loss': 0.3161, 'grad_norm': 0.1435796022415161, 'learning_rate': 0.00010105263157894738, 'epoch': 0.19}
{'loss': 0.3703, 'grad_norm': 0.14022934436798096, 'learning_rate': 9.894736842105263e-05, 'epoch': 0.19}
{'loss': 0.3592, 'grad_norm': 0.12194594740867615, 'learning_rate': 9.68421052631579e-05, 'epoch': 0.2}
{'loss': 0.3411, 'grad_norm': 0.11623945087194443, 'learning_rate': 9.473684210526316e-05, 'epoch': 0.2}
{'loss': 0.2982, 'grad_norm': 0.11288086324930191, 'learning_rate': 9.263157894736843e-05, 'epoch': 0.2}
{'loss': 0.3636, 'grad_norm': 0.19941461086273193, 'learning_rate': 9.052631578947369e-05, 'epoch': 0.21}
{'loss': 0.327, 'grad_norm': 0.11614833027124405, 'learning_rate': 8.842105263157894e-05, 'epoch': 0.21}
{'loss': 0.2447, 'grad_norm': 0.10045632719993591, 'learning_rate': 8.631578947368421e-05, 'epoch': 0.21}
{'loss': 0.3494, 'grad_norm': 0.11216063797473907, 'learning_rate': 8.421052631578948e-05, 'epoch': 0.22}
{'loss': 0.3233, 'grad_norm': 0.13398653268814087, 'learning_rate': 8.210526315789474e-05, 'epoch': 0.22}
{'loss': 0.2874, 'grad_norm': 0.11401170492172241, 'learning_rate': 8e-05, 'epoch': 0.22}
{'loss': 0.3704, 'grad_norm': 0.12382341921329498, 'learning_rate': 7.789473684210526e-05, 'epoch': 0.23}
{'loss': 0.3196, 'grad_norm': 0.1083361878991127, 'learning_rate': 7.578947368421054e-05, 'epoch': 0.23}
{'loss': 0.3459, 'grad_norm': 0.19833165407180786, 'learning_rate': 7.368421052631579e-05, 'epoch': 0.23}
{'loss': 0.3015, 'grad_norm': 0.10992985963821411, 'learning_rate': 7.157894736842105e-05, 'epoch': 0.24}
{'loss': 0.328, 'grad_norm': 0.11520015448331833, 'learning_rate': 6.947368421052632e-05, 'epoch': 0.24}
{'loss': 0.3837, 'grad_norm': 0.13615329563617706, 'learning_rate': 6.736842105263159e-05, 'epoch': 0.24}
{'loss': 0.3063, 'grad_norm': 0.12049426138401031, 'learning_rate': 6.526315789473685e-05, 'epoch': 0.25}
{'loss': 0.3201, 'grad_norm': 0.1054273247718811, 'learning_rate': 6.31578947368421e-05, 'epoch': 0.25}
{'loss': 0.2946, 'grad_norm': 0.11126892268657684, 'learning_rate': 6.105263157894737e-05, 'epoch': 0.26}
{'loss': 0.2455, 'grad_norm': 0.10475530475378036, 'learning_rate': 5.894736842105263e-05, 'epoch': 0.26}
{'loss': 0.335, 'grad_norm': 0.13587339222431183, 'learning_rate': 5.68421052631579e-05, 'epoch': 0.26}
{'loss': 0.3251, 'grad_norm': 0.10262706875801086, 'learning_rate': 5.4736842105263165e-05, 'epoch': 0.27}
{'loss': 0.2922, 'grad_norm': 0.1111547127366066, 'learning_rate': 5.2631578947368424e-05, 'epoch': 0.27}
{'loss': 0.3437, 'grad_norm': 0.12631115317344666, 'learning_rate': 5.052631578947369e-05, 'epoch': 0.27}
{'loss': 0.3183, 'grad_norm': 0.11674674600362778, 'learning_rate': 4.842105263157895e-05, 'epoch': 0.28}
100%|██████████| 100/100 [01:44<00:00,  1.04s/it]
{'loss': 0.3041, 'grad_norm': 0.1114904135465622, 'learning_rate': 4.6315789473684214e-05, 'epoch': 0.28}
{'loss': 0.3681, 'grad_norm': 0.11916472762823105, 'learning_rate': 4.421052631578947e-05, 'epoch': 0.28}
{'loss': 0.3591, 'grad_norm': 0.15139518678188324, 'learning_rate': 4.210526315789474e-05, 'epoch': 0.29}
{'loss': 0.2658, 'grad_norm': 0.11994612962007523, 'learning_rate': 4e-05, 'epoch': 0.29}
{'loss': 0.3452, 'grad_norm': 0.14532621204853058, 'learning_rate': 3.789473684210527e-05, 'epoch': 0.29}
{'loss': 0.3322, 'grad_norm': 0.12806951999664307, 'learning_rate': 3.578947368421053e-05, 'epoch': 0.3}
{'loss': 0.2921, 'grad_norm': 0.11126401275396347, 'learning_rate': 3.368421052631579e-05, 'epoch': 0.3}
{'loss': 0.2626, 'grad_norm': 0.10746820271015167, 'learning_rate': 3.157894736842105e-05, 'epoch': 0.3}
{'loss': 0.3157, 'grad_norm': 0.11800961941480637, 'learning_rate': 2.9473684210526314e-05, 'epoch': 0.31}
{'loss': 0.3579, 'grad_norm': 0.10837972909212112, 'learning_rate': 2.7368421052631583e-05, 'epoch': 0.31}
{'loss': 0.3326, 'grad_norm': 0.11892487853765488, 'learning_rate': 2.5263157894736845e-05, 'epoch': 0.32}
{'loss': 0.3138, 'grad_norm': 0.11013548076152802, 'learning_rate': 2.3157894736842107e-05, 'epoch': 0.32}
{'loss': 0.3079, 'grad_norm': 0.11282574385404587, 'learning_rate': 2.105263157894737e-05, 'epoch': 0.32}
{'loss': 0.3413, 'grad_norm': 0.11510809510946274, 'learning_rate': 1.8947368421052634e-05, 'epoch': 0.33}
{'loss': 0.3502, 'grad_norm': 0.12397557497024536, 'learning_rate': 1.6842105263157896e-05, 'epoch': 0.33}
{'loss': 0.4151, 'grad_norm': 0.12421295046806335, 'learning_rate': 1.4736842105263157e-05, 'epoch': 0.33}
{'loss': 0.3161, 'grad_norm': 0.12127997726202011, 'learning_rate': 1.2631578947368422e-05, 'epoch': 0.34}
{'loss': 0.2764, 'grad_norm': 0.1132940873503685, 'learning_rate': 1.0526315789473684e-05, 'epoch': 0.34}
{'loss': 0.3861, 'grad_norm': 0.12642629444599152, 'learning_rate': 8.421052631578948e-06, 'epoch': 0.34}
{'loss': 0.2874, 'grad_norm': 0.10199450701475143, 'learning_rate': 6.315789473684211e-06, 'epoch': 0.35}
{'loss': 0.269, 'grad_norm': 0.12433204054832458, 'learning_rate': 4.210526315789474e-06, 'epoch': 0.35}
{'loss': 0.3061, 'grad_norm': 0.11193855851888657, 'learning_rate': 2.105263157894737e-06, 'epoch': 0.35}
{'train_runtime': 104.4302, 'train_samples_per_second': 3.83, 'train_steps_per_second': 0.958, 'train_loss': 0.3433119258284569, 'epoch': 0.35}
|===========================================================================|
|                  PyTorch CUDA memory summary, device ID 0                 |
|---------------------------------------------------------------------------|
|            CUDA OOMs: 0            |        cudaMalloc retries: 0         |
|===========================================================================|
|        Metric         | Cur Usage  | Peak Usage | Tot Alloc  | Tot Freed  |
|---------------------------------------------------------------------------|
| Allocated memory      |  19993 MiB |  27090 MiB |  21744 GiB |  21725 GiB |
|       from large pool |  19480 MiB |  26746 MiB |  21552 GiB |  21533 GiB |
|       from small pool |    513 MiB |    688 MiB |    192 GiB |    191 GiB |
|---------------------------------------------------------------------------|
| Active memory         |  19993 MiB |  27090 MiB |  21744 GiB |  21725 GiB |
|       from large pool |  19480 MiB |  26746 MiB |  21552 GiB |  21533 GiB |
|       from small pool |    513 MiB |    688 MiB |    192 GiB |    191 GiB |
|---------------------------------------------------------------------------|
| Requested memory      |  19930 MiB |  26912 MiB |  21730 GiB |  21711 GiB |
|       from large pool |  19417 MiB |  26567 MiB |  21538 GiB |  21519 GiB |
|       from small pool |    513 MiB |    688 MiB |    191 GiB |    191 GiB |
|---------------------------------------------------------------------------|
| GPU reserved memory   |  25634 MiB |  27320 MiB |  35158 MiB |   9524 MiB |
|       from large pool |  24920 MiB |  26960 MiB |  34220 MiB |   9300 MiB |
|       from small pool |    714 MiB |    714 MiB |    938 MiB |    224 MiB |
|---------------------------------------------------------------------------|
| Non-releasable memory |      0 B   |      0 B   |      0 B   |      0 B   |
|       from large pool |      0 B   |      0 B   |      0 B   |      0 B   |
|       from small pool |      0 B   |      0 B   |      0 B   |      0 B   |
|---------------------------------------------------------------------------|
| Allocations           |    4232    |    5249    |    1053 K  |    1049 K  |
|       from large pool |     332    |     457    |     393 K  |     392 K  |
|       from small pool |    3900    |    4809    |     660 K  |     656 K  |
|---------------------------------------------------------------------------|
| Active allocs         |    4232    |    5249    |    1053 K  |    1049 K  |
|       from large pool |     332    |     457    |     393 K  |     392 K  |
|       from small pool |    3900    |    4809    |     660 K  |     656 K  |
|---------------------------------------------------------------------------|
| GPU reserved segments |       0    |       0    |       0    |       0    |
|       from large pool |       0    |       0    |       0    |       0    |
|       from small pool |       0    |       0    |       0    |       0    |
|---------------------------------------------------------------------------|
| Non-releasable allocs |       0    |       0    |       0    |       0    |
|       from large pool |       0    |       0    |       0    |       0    |
|       from small pool |       0    |       0    |       0    |       0    |
|---------------------------------------------------------------------------|
| Oversize allocations  |       0    |       0    |       0    |       0    |
|---------------------------------------------------------------------------|
| Oversize GPU segments |       0    |       0    |       0    |       0    |
|===========================================================================|

You are given a problem.
Think about the problem and provide your working out.
Place it between <start_working_out> and <end_working_out>.
Then, provide your solution between <SOLUTION></SOLUTION><|im_end|>Given $\sqrt{x^2+165}-\sqrt{x^2-52}=7$ and $x$ is positive, find all possible values of [rank0]: Traceback (most recent call last):
[rank0]:   File "/home/andrewor/scratch/datta_test_script.py", line 183, in <module>
[rank0]:     _ = model.generate(
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/peft/peft_model.py", line 1973, in generate
[rank0]:     outputs = self.base_model.generate(*args, **kwargs)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/models/llama.py", line 2036, in unsloth_fast_generate
[rank0]:     output = self._old_generate(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 120, in decorate_context
[rank0]:     return func(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/transformers/generation/utils.py", line 2539, in generate
[rank0]:     result = self._sample(
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/transformers/generation/utils.py", line 2867, in _sample
[rank0]:     outputs = self(**model_inputs, return_dict=True)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
[rank0]:     return self._call_impl(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
[rank0]:     return forward_call(*args, **kwargs)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/models/llama.py", line 1327, in _CausalLM_fast_forward
[rank0]:     outputs = self.model(
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
[rank0]:     return self._call_impl(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
[rank0]:     return forward_call(*args, **kwargs)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/models/llama.py", line 1105, in LlamaModel_fast_forward
[rank0]:     layer_outputs = decoder_layer(
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/transformers/modeling_layers.py", line 94, in __call__
[rank0]:     return super().__call__(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
[rank0]:     return self._call_impl(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
[rank0]:     return forward_call(*args, **kwargs)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/models/llama.py", line 728, in LlamaDecoderLayer_fast_forward
[rank0]:     hidden_states, self_attn_weights, present_key_value = self.self_attn(
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
[rank0]:     return self._call_impl(*args, **kwargs)
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
[rank0]:     return forward_call(*args, **kwargs)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/models/qwen3.py", line 90, in Qwen3Attention_fast_forward
[rank0]:     Q, K, V = self.apply_qkv(self, hidden_states)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/kernels/fast_lora.py", line 527, in apply_lora_qkv
[rank0]:     Q, K, V = LoRA_QKV.apply(
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/autograd/function.py", line 581, in apply
[rank0]:     return super().apply(*args, **kwargs)  # type: ignore[misc]
[rank0]:   File "/home/andrewor/.conda/envs/pytorch-3.10/lib/python3.10/site-packages/torch/amp/autocast_mode.py", line 527, in decorate_fwd
[rank0]:     return fwd(*args, **kwargs)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/kernels/fast_lora.py", line 382, in forward
[rank0]:     Q = matmul_lora(X, QW, QW_quant, QA, QB, QS)
[rank0]:   File "/home/andrewor/local/unsloth/unsloth/kernels/utils.py", line 1019, in matmul_lora
[rank0]:     out = torch_matmul(X, W.t(), out = out)
[rank0]:   File "/home/andrewor/local/ao/torchao/utils.py", line 609, in _dispatch__torch_function__
[rank0]:     return func(*args, **kwargs)
[rank0]: RuntimeError: Cannot set version_counter for inference tensor
[rank0]:[W1118 14:15:29.020192022 ProcessGroupNCCL.cpp:1524] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())