[{"data":1,"prerenderedAt":676},["ShallowReactive",2],{"project-39261486":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":35,"aspects":36,"strip":47,"post_text":48,"what":49,"said":82,"maker_in_thread":171,"shape":183,"grid":187,"attempts":638,"same_week":639,"maker_shows":43,"maker_hidden":668,"reader_version":669,"card":670},"39261486","Natural-SQL-7B","A strong text-to-SQL model","title","Show HN: Natural-SQL-7B, a strong text-to-SQL model","https:\u002F\u002Fgithub.com\u002Fcfahlgren1\u002Fnatural-sql","github.com","thecalebf",1707142956,2024,362,168,134,3515,4,"legend","ai","text-to-sql model","library",[24,25,26,27],"text-to-sql","natural language sql","sql generation model","local llm","mixed",-0.162,{"positive":31,"neutral":32,"mixed":33,"critical":34},10,21,2,15,48,{"performance":37,"privacy":39,"execution":41,"idea":44,"docs":46},{"positive":38,"critical":18},1,{"positive":40,"critical":40},0,{"positive":42,"critical":43},5,3,{"positive":43,"critical":45},7,{"positive":40,"critical":40},[43,40,40,40,38,40,40,40,40,40,40,40,40,40,40,40,40,40,40,38,33,42,42,38,42,38,40,43,43,43,43,42,38,43,43,38,38,43,43,38],"Would love thoughts!\n\nHere is the HF page: https:\u002F\u002Fhuggingface.co\u002Fchatdb\u002Fnatural-sql-7b",{"problem":50,"audience":51,"pricing":50,"motivation":50,"time_to_build":50,"stack":62,"openness":80,"stage":81,"team_size":80},null,{"value":52,"source":53,"evidence":54},"Non-technical users and analysts who need SQL answers","inferred",[55,59],{"id":56,"by":57,"text":58},"39262883","croes","I doubt it's useful for complex queries or databases without proper relation info in the database schemes. So it's limited to rather simple queries for users without SQL…",{"id":60,"by":11,"text":61},"39262979","Yes it is designed for users without SQL knowledge, however, it can still perform fairly well with questions on the difficult side (for non technical users) with queries having…",[63,74],{"value":64,"source":65,"evidence":66},"PostgreSQL (evaluation and training target)","maker",[67,71],{"id":68,"by":69,"text":70},"39289514","moltar","The card states: > This model was evaluated on SQL-Eval, a PostgreSQL-based evaluation framework developed by Defog for testing and alignment of model capabilities. But this…",{"id":72,"by":11,"text":73},"39262351","No, those are benchmark, evaluation questions. The fine tune dataset was a custom, synthetically generated dataset of ~20k PostgreSQL Text to SQL pairs covering different SQL…",{"value":75,"source":65,"evidence":76},"Fine-tuned from DeepSeek-Coder",[77],{"id":78,"by":11,"text":79},"39263511","Sorry, it seems complicated. Since it is a finetune of Deepseek-Coder, I had to include their license. Deepseek is pretty open, just says not to use it for: - military purposes -…","unknown","beta",{"praise":83,"objections":96,"requests":123,"questions":139,"alternatives":154},[84],{"value":85,"source":86,"evidence":87},"Useful as a first draft for SQL queries","commenter",[88,92],{"id":89,"by":90,"text":91},"39262450","rgbrgb","So it looks like it scores 76.5% on SQL-Eval [0], a bit behind GPT-4 at 83% and sqlcoder-15b at 78%. What kind of applications would this be useful for? What can you build with…",{"id":93,"by":94,"text":95},"39270932","zeroq","For anyone who says it's useless because it's only 75% correct, please consider these two points: (1) this is the first instalment, and it's already close to be a thousand times…",[97,101,112],{"value":98,"source":86,"evidence":99},"Only about 76.5% accurate on SQL-Eval, behind larger models",[100],{"id":89,"by":90,"text":91},{"value":102,"source":86,"evidence":103},"Does not understand business meaning of column data",[104,108],{"id":105,"by":106,"text":107},"39266224","lolpanda","I don't think any of those text-to-sql models are solving the right problems. The hard part is not syntax or I don't know how to write a group by query. Most data scientists and…",{"id":109,"by":110,"text":111},"39269514","ainesh93","Hard to claim success with \"complex\" questions if you don't account for business context and organizational nuances. For example, \"active\" listings on Redfin may be a combination…",{"value":113,"source":86,"evidence":114},"Custom license with use restrictions is not open source",[115,119],{"id":116,"by":117,"text":118},"39262180","satvikpendem","This is not open source, because you have use-based restrictions. Call it what it is, source available.",{"id":120,"by":121,"text":122},"39263375","wantsanagent","So I need to ask my lawyer to review a custom license before I can consider using this? No thanks.",[124,132],{"value":125,"source":86,"evidence":126},"Support for other SQL dialects such as BigQuery, MySQL, DuckDB",[127,131],{"id":128,"by":129,"text":130},"39268311","floridageorgia","I would 100% use this if you had a text -> Big Query SQL feature. Please add.",{"id":68,"by":69,"text":70},{"value":133,"source":86,"evidence":134},"Support for longer context windows for large schemas",[135],{"id":136,"by":137,"text":138},"39262312","xfalcox","Context is 4096? My app db DDL is 19877 tokens (using Llama2 tokenizer) long, so that means we need to do a RAG for handling the DDL prompt injection. A model like this with a…",[140,147],{"value":141,"source":86,"evidence":142},"Can it be used with Vanna or other tools under its license",[143],{"id":144,"by":145,"text":146},"39263199","zainhoda","Very cool. Would the license allow for use with Vanna? https:\u002F\u002Fgithub.com\u002Fvanna-ai\u002Fvanna",{"value":148,"source":86,"evidence":149},"How was the model fine-tuned",[150],{"id":151,"by":152,"text":153},"39264140","htrp","Would be very interesting if you wrote a bit about what you did to actually do the finetuning",[155,163,167],{"value":156,"source":86,"evidence":157},"GPT-4 and ChatGPT",[158,159],{"id":89,"by":90,"text":91},{"id":160,"by":161,"text":162},"39268665","aussieguy1234","I've been using GPT-4 to write SQL to get a bunch of insights. Its very good at writing up queries for all kinds of metrics. These queries answer questions that would have…",{"value":164,"source":86,"evidence":165},"sqlcoder-15b",[166],{"id":89,"by":90,"text":91},{"value":168,"source":86,"evidence":169},"Vanna",[170],{"id":144,"by":145,"text":146},{"comments":172,"answered":45,"key":173},12,[174,177,180],{"id":175,"text":176},"39262289","Updated the title since it may have been confusing, appreciate the feedback!",{"id":178,"text":179},"39262260","Sure, there is a blog post with some other examples of the first model iteration with more complex, multipart questions.…",{"id":181,"text":182},"39262164","Thanks! Yes, that was an earlier iteration. The v1 is here https:\u002F\u002Fhuggingface.co\u002Fchatdb\u002Fnatural-sql-7b. Plan to push to Ollama soon and build an open source free tool around it.…",{"commenters":184,"max_depth":31,"maker_replies":172,"first_comment_mins":185,"lifespan_hours":186},97,58,49,[188,190,194,198,201,205,208,210,213,215,217,220,222,225,227,229,232,235,237,240,242,244,247,251,254,257,259,262,264,266,269,271,274,276,279,282,284,287,289,292,295,298,300,302,305,307,309,311,313,316,318,320,322,324,327,329,331,333,335,337,340,342,345,348,350,352,355,358,360,363,366,369,372,374,377,379,381,384,386,389,391,394,397,400,404,407,410,414,418,420,424,427,431,433,436,438,441,442,445,448,452,454,456,459,461,464,465,467,469,472,475,477,480,482,485,488,489,492,495,499,502,505,509,512,516,517,519,523,526,528,531,534,537,539,540,542,544,548,550,553,556,559,562,566,570,573,577,578,581,583,585,587,590,592,595,597,601,604,608,611,613,617,621,625,627,630,631,634],[89,40,43,189,90,40],"project\u002Fneutral\u002Fperformance",[191,40,40,192,193,40],"39262512","commenter\u002Fneutral\u002Fprivacy","whimsicalism",[195,40,40,196,197,40],"39262729","other\u002Fneutral\u002Fprivacy","tempusalaria",[199,40,40,196,200,40],"39262866","anonylizard",[202,40,40,203,204,40],"39262928","commenter\u002Fneutral\u002Fnone","CharlesW",[206,40,40,207,200,40],"39263491","other\u002Fneutral\u002Fnone",[209,40,40,203,204,40],"39263600",[211,40,40,207,212,40],"39268350","drittich",[214,40,40,196,197,40],"39263431",[216,40,40,203,193,40],"39263590",[218,40,40,196,219,40],"39263551","ParetoOptimal",[221,40,40,192,193,40],"39262875",[223,40,40,196,224,40],"39263291","Nevermark",[226,40,40,192,193,40],"39263359",[228,40,40,196,224,40],"39263583",[230,40,40,196,231,40],"39270424","hackerlight",[233,40,38,234,200,40],"39262739","project\u002Fcritical\u002Fperformance",[236,40,40,203,193,40],"39262846",[238,40,43,239,90,40],"39263054","project\u002Fneutral\u002Fprivacy",[241,40,40,196,90,40],"39262767",[243,40,40,196,193,40],"39262907",[245,40,40,196,246,40],"39264276","ren_engineer",[248,40,40,249,250,40],"39262921","commenter\u002Fcritical\u002Fperformance","fifilura",[252,40,40,203,253,40],"39263145","gavinray",[255,40,40,203,256,40],"39263201","neuronexmachina",[258,40,40,203,253,40],"39263305",[260,40,40,203,261,40],"39263445","ttfkam",[263,40,40,203,253,40],"39263559",[265,40,40,203,250,40],"39263805",[267,40,40,203,268,40],"39266452","koolba",[270,40,40,203,261,40],"39266361",[272,40,40,207,273,40],"39263455","samstave",[275,40,40,203,250,40],"39263310",[277,40,40,203,278,40],"39264360","3rd3",[280,40,40,203,281,40],"39265926","ComputerGuru",[283,40,40,203,278,40],"39266026",[285,40,40,203,286,40],"39266464","ako",[288,40,40,203,278,40],"39266790",[290,40,40,203,291,40],"39270113","ses1984",[293,40,40,203,294,40],"39265812","beefield",[296,40,40,207,297,40],"39269251","fijiaarone",[299,40,40,203,250,40],"39271550",[301,40,40,203,286,40],"39271647",[303,40,40,203,304,40],"39281624","riku_iki",[306,40,40,207,281,40],"39265904",[308,40,40,207,250,40],"39265998",[310,40,40,203,281,40],"39266278",[312,40,40,203,90,40],"39263184",[314,40,40,203,315,40],"39263686","calvinmorrison",[317,40,40,203,250,40],"39263908",[319,40,40,203,315,40],"39265186",[321,40,40,203,250,40],"39265859",[323,40,40,203,286,40],"39266590",[325,40,40,203,326,40],"39267643","threeseed",[328,40,40,203,250,40],"39267907",[330,40,40,203,326,40],"39268231",[332,40,40,203,250,40],"39271169",[334,40,40,203,286,40],"39271574",[336,40,40,207,286,40],"39264635",[338,40,40,207,339,40],"39265174","brian_reardon",[341,40,40,207,286,40],"39272043",[343,40,40,207,344,40],"39264327","brightball",[346,40,40,207,347,40],"39267379","pavlov",[349,40,40,203,250,40],"39268318",[351,40,40,207,326,40],"39267491",[353,40,40,203,354,40],"39269632","tomnipotent",[356,40,40,203,357,40],"39270329","hobs",[359,40,40,203,354,40],"39270615",[361,40,40,207,362,40],"39267743","panarky",[364,40,40,203,365,40],"39284263","ecocentrik",[367,40,40,207,368,40],"39268302","miohtama",[370,40,40,207,371,40],"39268666","HideousKojima",[373,40,40,207,368,40],"39268952",[375,40,40,207,376,40],"39269091","munk-a",[378,40,40,207,286,40],"39271608",[380,40,40,207,354,40],"39269596",[382,40,40,207,383,40],"39273141","freilanzer",[385,40,40,207,161,40],"39268704",[387,40,40,207,388,40],"39269027","abhibeckert",[390,40,40,203,161,40],"39269975",[392,40,40,207,393,40],"39276535","klibertp",[395,40,40,207,396,40],"39273989","pid-1",[398,40,40,207,399,40],"39274065","vmfunction",[401,40,38,402,403,40],"39263371","project\u002Fcritical\u002Fexecution","wg0",[405,40,40,192,406,40],"39263820","bongodongobob",[408,40,38,402,409,40],"39263822","internet101010",[411,40,42,412,413,40],"39263982","project\u002Fpositive\u002Fexecution","Closi",[415,40,38,416,417,40],"39264131","project\u002Fcritical\u002Fidea","tillvz",[419,40,42,412,413,40],"39263865",[421,40,42,422,423,40],"39264405","project\u002Fpositive\u002Fidea","davidy123",[425,40,43,426,339,40],"39265148","project\u002Fneutral\u002Fexecution",[428,40,43,429,430,40],"39265674","project\u002Fneutral\u002Fidea","dimask",[432,40,42,412,294,40],"39265782",[434,40,42,412,435,40],"39266513","bagels",[437,40,42,412,297,40],"39269279",[439,40,38,234,440,40],"39269483","lawxls",[105,38,38,416,106,40],[443,38,40,207,444,40],"39266374","joshhart",[446,38,40,207,447,40],"39268233","l5870uoo9y",[449,38,40,450,451,40],"39266704","other\u002Fneutral\u002Fidea","mritchie712",[453,38,40,450,447,40],"39268207",[455,38,40,450,451,40],"39276776",[457,38,43,429,458,40],"39266987","danielmarkbruce",[460,38,40,450,447,40],"39268114",[462,38,42,422,463,40],"39270136","edmundsauto",[116,33,38,416,117,40],[175,33,40,466,11,38],"maker\u002Fneutral\u002Fnone",[468,33,38,416,204,40],"39262942",[470,33,40,203,471,40],"39262971","Keyframe",[473,33,40,207,474,40],"39266300","klabb3",[476,33,40,207,471,40],"39266841",[478,33,40,203,479,40],"39272727","Palmik",[481,33,40,203,117,40],"39274701",[483,33,40,207,484,40],"39266862","gwd",[486,43,38,402,487,40],"39262231","itsoktocry",[178,43,40,466,11,38],[490,43,42,491,487,40],"39262934","maker\u002Fpositive\u002Fnone",[493,43,40,207,494,40],"39266163","int_19h",[496,18,43,497,498,40],"39264423","project\u002Fmixed\u002Fexecution","buzzm",[500,18,43,429,501,40],"39264861","owlstuffing",[503,42,43,189,504,40],"39262493","CastFX",[506,507,40,450,508,40],"39263342",6,"jimmytucson",[510,507,40,207,511,40],"39264231","fwip",[513,45,43,514,515,40],"39262130","project\u002Fmixed\u002Fnone","zurfer",[181,45,40,491,11,38],[518,45,40,203,515,40],"39262368",[520,45,40,521,522,40],"39262935","commenter\u002Fpositive\u002Fnone","eurekin",[144,524,43,525,145,40],8,"project\u002Fneutral\u002Fnone",[527,524,40,491,11,38],"39263527",[529,524,43,525,530,40],"39264027","pama",[93,532,42,533,94,40],9,"project\u002Fpositive\u002Fperformance",[535,31,40,450,536,40],"39263148","Tycho",[68,538,43,525,69,40],11,[160,172,40,207,161,40],[136,541,38,234,137,40],13,[543,541,40,491,11,38],"39262364",[545,546,38,416,547,40],"39262432",14,"cuuupid",[549,546,40,466,11,38],"39262465",[551,546,40,552,417,40],"39263109","commenter\u002Fneutral\u002Fidea",[554,34,38,416,555,40],"39264351","K0IN",[557,34,40,203,558,40],"39265295","bm-rf",[560,561,42,422,501,40],"39264569",16,[563,564,43,565,69,40],"39289508",17,"project\u002Fneutral\u002Fdocs",[567,568,43,429,569,40],"39262186",18,"delichon",[571,568,40,572,11,38],"39262389","maker\u002Fpositive\u002Fidea",[574,575,43,525,576,40],"39262329",19,"bottlepalm",[72,575,40,466,11,38],[579,575,43,525,580,40],"39262441","Semaphor",[582,575,40,466,11,38],"39262496",[56,584,38,234,57,40],20,[60,584,40,586,11,38],"maker\u002Fpositive\u002Fperformance",[588,584,40,203,589,40],"39262992","lofties",[591,584,40,203,57,40],"39269458",[593,32,43,565,594,40],"39262110","mrgaro",[109,596,38,416,110,40],22,[598,599,40,207,600,40],"39267028",23,"Uptrenda",[602,599,40,207,603,40],"39271037","jamil7",[605,606,43,525,607,40],"39270985",24,"Too",[151,609,43,610,152,40],25,"maker\u002Fneutral\u002Fdocs",[128,612,43,525,129,40],26,[614,615,40,207,616,40],"39270475",27,"brendongeils",[618,619,43,429,620,40],"39264598",28,"paidcompute",[622,623,43,525,624,40],"39262213",29,"swalsh",[626,623,40,466,11,38],"39262270",[120,628,38,629,121,40],30,"project\u002Fcritical\u002Fnone",[78,628,40,466,11,38],[632,633,40,207,297,40],"39269199",31,[635,636,43,525,637,40],"39282131",32,"saigal",[],[640,645,650,655,660,664],{"id":641,"name":642,"points":643,"category":644},"39288039","One Minute Focus",741,"consumer_social",{"id":646,"name":647,"points":648,"category":649},"39263854","Atopile",595,"dev_tools",{"id":651,"name":652,"points":653,"category":654},"39214743","filippo.io\u002Fmlkem768",315,"languages",{"id":656,"name":657,"points":658,"category":659},"39232572","Minesweeper Twist",293,"games",{"id":661,"name":662,"points":663,"category":644},"39260760","Random Hacker News",267,{"id":665,"name":666,"points":667,"category":659},"39268439","Name That Nation",201,false,"v2",{"id":671,"name":5,"line":6,"ai":668,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":672,"top":18,"kw":673,"maker":11,"domain":10,"tier":19,"attempts":40,"strip":674,"url":9,"mix":675,"praise":85,"objection":98,"sub":21},39261486,-0.16,[24,25,26,27],"3000100000000000000125515103333513311331",[31,32,33,34],1794036649243]