-
Apparently, I cannot use a function such as `1/(1 + exp(-x))` in the contrast command.
``` set.seed(1)
age
sdaza updated
8 years ago
-
### What happened + What you expected to happen
The problem is here:
https://github.com/ray-project/ray/blob/6d8d7398df4f90abd008468c5b4fb1ebfa587256/rllib/models/tf/tf_action_dist.py#L90-L91
[tf…
-
Hi,
Is there a way to change the frequency_penality or logit bias when sending a completion request?
-
I encountered an issue while finetune with the officially released code using the DeepSpeed. Here is the detailed error message:
```
File "/lib/python3.11/site-packages/deepspeed/runtime/zero/linear…
-
Exp
Logit
Identity
Through family() interface as well
-
when i train"train_generation_model",i meet the problem
Traceback (most recent call last):
File "./train_generative_model.py", line 160, in
g_train, d_train, sampler, saver, loader, extras = get_mo…
-
**Is your feature request related to a problem? Please describe.**
While HuggingFace is already supported through the `OpenAICall` class (use base_url, model, api_key="-"), we should support HuggingF…
-
Hello, I would like to ask about the meaning of tokens being integers. I noticed that the final forward pass to the tokenizer involves the `cls_logits_softmax` tensor, and it directly performs a matri…
-
### 📚 The doc issue
currently distributed collective docs don't mention functional collectives.
PT2 support needs functional collectives to be used. (also, PT2 supports transparently rewriting _S…
-
Thank you for this nice paper, your new insights and the detailed Training setup description in Section 3.1.
You mention that you are using PyTorch FSDP for training. I have some additional questi…