Turbo Hang

[Roxine Denison]

Ihave updated my code from gpt-3.5-turbo-0613 to gpt-3.5-turbo-1106, but the code seems to hang for no reason. When I first tested it, it ran without issue and actually worked out faster that the older model.

Are you using any kind of VPN or Proxy? What server kind of hosting infrastructure are you using? Is it on Azure, AWS, Google, Commercial VPS, home internet? Can you post a code snippet of the API calling code along with any setup it relies on, please?

The server infrastructure is under heavy load with the all of the new users joining, hence the pause placed on new Plus memberships, this should get better over the coming days. Most of the lack of responses and hangs at the moment are related to this.

I switched from gpt-3.5-turbo-1106 to gpt-4-1106-preview and that helped quite a bit with the hanging, but at 10X the price, I hope they resolve this issue soon so I can switch back to the cheaper model.

I know I have suggested many solutions, but if anyone lands here again. I have found work around with a timeout function, if the GPT3.5 call hangs for over 20sec, I call GPT4. So far I have been able to catch the hanged calls and at least respond within 20sec - then mix between the cheaper and faster GPT3.5 turbo and more expensive GPT4. This way I can even count how many times a day the function hangs.

Here is my workaround . . .

If the API is failing 1 in 10 times (as has been typical for gpt-3.5-turbo-1106 if not even worse) you can dispatch the answer to two parallel calls of the model at the same time and still have even lower expense, especially if streaming and closing the connection on the one that is not first to respond.

Hey! It turns out there was a bug on our end that could result in timeouts in certain scenarios. We have since fixed the issue. Please let us know in a new thread if you end up seeing similar issues again. Thanks again for reporting this!

Using a standard python request to call the API typically performs quite well, however randomly the request will hang indefinitely (been >2 hours and no response) - have anyone experienced a similar problem and if so how did you solve it?

I too am on this page (however many months later) cause requests to gpt-3.5-turbo-1106 frequently seem to stall today. Worked lightning fast earlier yesterday, but began having issues during the evening. The 4-preview API still seems to work fine.

I had the same issue and solved it with the optional parameter request_timeout and incremental timeout times. The approach is like this. This pattern is quite common for REST requests as timeouts may often occur because of overloads.

Picking a bear hang starts with finding a tree with an appropriate branch. Fifteen feet up and five feet from the base of the tree is the minimum required, but with ursine acrobats, the higher the hang, the better. If you can hang your food from a limb thirty to fifty feet high, it's much more risky for a bear to attempt a grab. It also gives you the option of placing your food bag ten to fifteen feet below your hanging branch, which further confuses bears. If possible, pick a limb that really sticks out, one that is six inches in diameter at the attachment to the tree and has multiple branches - preferably that obstruct a bear's progress. It's best if your branch stands somewhat alone - it's easier to keep the food away from other branches the bear could use, and helps keep your food from getting stuck when retrieving it. It's important to pick a branch too thick for a bear to break off but to hang your rope far enough out on this branch that it won't support a bear's weight - even a cub's. Ponderosa pines and ancient lodgepoles most frequently have branches that meet these criteria. Pick a live tree and branch for your hang, because dead wood, even when very thick, is much more brittle and easy for a bear to break. The rope is less likely to slip off of your branch if the branch tip turns up or if there are perpendicular side branches to catch the rope. You don't want your rope to slide off of the branch should your adversary bend it while walking part way out.

Really well-written and comprehensive article; nice video, too! However, since I'm only a 5-foot-tall woman, I really think I'll work on lightening my pack weight as much as possible so I can use a bear canister. I think it would be much easier for me.

Interesting, but if the author is still bear bagging in Yosemite or the JMT corridor, he is risking a fine and being escorted out of the wilderness by a ranger. Bear canisters on the SIBBG approved list are mandatory, even if you turbo hang.

This is a helpful article, but Kevin should have noted that in hanging his food in Tuolumne Meadows, he violated Park Service policy. In all of Yosemite, the use of bear canisters (or food storage lockers) is mandatory.

I've never been able to hit the side of a barn when throwing anything. The jerky motions required for throwing also start up painful arthritis in my shoulders. I therefore cannot hang my food. It's also almost impossible to find appropriate trees near or above timberline where I prefer to backpack. I therefore use an Ursack with an OP Sack liner. The Ursack also keeps squirrels and birds out of my food, which a lightweight stuff sack does not. If legally required, I use a canister. In the Pacific Northwest, that's in Olympic National Park.

"Interesting, but if the author is still bear bagging in Yosemite or the JMT corridor, he is risking a fine and being escorted out of the wilderness by a ranger. Bear canisters on the SIBBG approved list are mandatory, even if you turbo hang.

and

This is a helpful article, but Kevin should have noted that in hanging his food in Tuolumne Meadows, he violated Park Service policy. In all of Yosemite, the use of bear canisters (or food storage lockers) is mandatory"

This is true, Kevin, but with one caveat: The frontcountry types, at least in the Owens Valley Ranger Stations, who actually issue the permits go through a spiel about proper hanging technique, which holds that there is only one permissible way to hang your food, i.e. the old fashioned counter balance method. They even give you a sheet demonstrating same and require you to initial a line on your permit stating that you have read it. I guess this is so you cannot plead ignorance if they catch you using another method. They insist that you follow it if you are not carrying a canister. I tried to reason with one of them about the weaknesses of the old counter balance method and she really got in my face about it. I got the distinct impression she was about to hold up my permit if I didn't back off. A typical front country bureaucrat, but they are the ones issuing the permits. My point is to be discreet if you use another method in SEKI or the wilderness areas on its eastern border, especially where you are likely to encounter a ranger. That said, I use the PCT method anyway, but I frequent areas where running into a ranger is not high on my list of concerns, nor bears for that matter. However, you have just given me a very interesting alternative to the PCT method, which I find problematic because I am only 5" 8" and often find it difficult to tie off the clove hitch high enough to keep the food bag 10' off the ground. I take comfort by telling myself that 99.9% of Sierra bears can't reach 9' off the ground, which is about the max that I can usually achieve. I'm already thinking of switching to your method, though, as it seems easier and more effective. Thanks for putting this article on BPL.

Nice article with video. Now come East and try that cowboy toss in a dense forest. : ) Good thing to know. As someone else mentioned, being short and female, I too have lightened my load enough to carry a small bear canister and not deal with all that, especially when I'm tired. And while all my other hiking friends are busy finding a suitable tree, I sit on the canister and watch the entertainment.

Also you get throws to 50 feet if you flake your p-cord on clear ground or your tarp/tent fabric and throw underhand with about a foot of slack or so. Even better is if you have a plastic or metal ring that you can put a loop of cord through and pinch then throw that. Practice makes perfect.

This is standard tree climbing stuff that arborists, recreational tree climbers, and canopy scientists use to get into short trees or inside crowns of tall trees: first they throw p-cord with a throw bag.

3a8082e126