New data and scripts

ass_extract.ipynb

@@ -1,90 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import ass\n",
-    "import os"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 6,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "SUB_NAME = \"2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].ass\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "cn_name = \"Extracted - \" + os.path.splitext(SUB_NAME)[0] + \".cn.txt\"\n",
-    "en_name = \"Extracted - \" + os.path.splitext(SUB_NAME)[0] + \".en.txt\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "with open(SUB_NAME, \"r\", encoding='utf-8-sig') as f:\n",
-    "    doc = ass.parse(f)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "cn_file = open(cn_name, \"w\")\n",
-    "en_file = open(en_name, \"w\")\n",
-    "\n",
-    "for e in doc.events:\n",
-    "    if e.style == '英':\n",
-    "        en_file.write(f\"[{e.start.total_seconds()}] {e.text.strip()} [{e.end.total_seconds()}]\\n\")\n",
-    "    elif e.style == '中':\n",
-    "        cn_file.write(f\"[{e.start.total_seconds()}] {e.text.strip()} [{e.end.total_seconds()}]\\n\")\n",
-    "\n",
-    "cn_file.close()\n",
-    "en_file.close()"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "whisper",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.10.13"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}

ass_extract.py

@@ -0,0 +1,23 @@
+import ass
+import os
+import sys
+
+SUB_NAME = sys.argv[1]
+
+cn_name = "Extracted - " + os.path.splitext(SUB_NAME)[0] + ".cn.txt"
+en_name = "Extracted - " + os.path.splitext(SUB_NAME)[0] + ".en.txt"
+
+with open(SUB_NAME, "r", encoding='utf-8-sig') as f:
+    doc = ass.parse(f)
+
+cn_file = open(cn_name, "w")
+en_file = open(en_name, "w")
+
+for e in doc.events:
+    if e.style == '英':
+        en_file.write(f"[{e.start.total_seconds()}] {e.text.strip()} [{e.end.total_seconds()}]\n")
+    elif e.style == '中':
+        cn_file.write(f"[{e.start.total_seconds()}] {e.text.strip()} [{e.end.total_seconds()}]\n")
+
+cn_file.close()
+en_file.close()

chatgpt-recent-test.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dfa03d5d56228528cb8858e5745a09ef245a711a991801ecbe6df49dc61213f8
-size 66253
+oid sha256:8c721a96f67922048c259225c859fae383021412513bbe299f38e86d103a0ebf
+size 61114

chatgpt-recent-train.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f96b1cbe7a65c9fe1b0ba02eff14823428c9b96f58cf0e68518919bd5055477
-size 269032
+oid sha256:ac5a43f261f248c2767ead6cb3a779a44366c38832648ed7ab6cccbede8f2ed7
+size 273326

chatgpt-test.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c95e22b1989640da79ffda2446cb6b199485fd220e71bfa289cdbdd9d7d38c6
-size 136957
+oid sha256:4b47b5bf70383550041c1502437cedfd19e7927980e76eead28f9a92887e07f0
+size 148931

chatgpt-train.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd2d9dc5501adf649acaa455e8cbc5289733fdda4aa8735af63623be06b58772
-size 576377
+oid sha256:6922bb916006ab705c7e83d0beda277afe23ee832e1380e59a5d1a777655ff8b
+size 621613

data/2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].cn.txt

@@ -100,7 +100,7 @@
 [290.66] 因为 大并不一定就是好 [292.42]
 [293.04] 适合任何场合 尤其是狭小空间 [295.24]
 [295.5] 小短粗能完美适配 [296.82]
-[296.82] 棘轮机构让拧螺丝变得更♂有♀乐趣 [300.06]
+[296.82] 棘轮机构让拧螺丝变得更有乐趣 [300.06]
 [300.68] 哦 Dennis [301.52]
 [302.06] 只用两根手指 你就可以轻松拧紧螺丝 [304.54]
 [304.68] 你可以把小短粗放在任何地方 [306.26]

data/2023-11-06 Grok [rCmsoa2Zlhg].cn.txt

@@ -1,5 +1,5 @@
 [0.44] 新的一周 新的科技新闻 [3.08]
-[3.08] 说来～ [3.36]
+[3.08] 说来 [3.36]
 [3.58] 这让我想起了我在计科系学的一首歌 [6.18]
 [6.78] 哗～开关一拨它就嗖～ [9.38]
 [9.38] 埃隆·马斯克的人工智能公司 嗯 还能叫什么呢 [12.14]

data/2023-11-13 The EU Will Break Apple. [Hod69Y1aVlY].cn.txt

@@ -0,0 +1,152 @@
+[0.0] 啊 您来这里是想看科技新闻 [3.76]
+[3.96] 我明白的 [5.4]
+[5.92] 所以我已经准备好了 [10.66]
+[10.66] 就在这儿 稍等 [12.42]
+[12.84] 苹果不希望iPhone用户 [16.06]
+[16.06] 从Appstore之外的地方下载安装应用 [18.12]
+[18.12] 也就是俗称的"通过外部源安装" [19.54]
+[19.54] 但根据Mark Gurman的报道 该公司似乎 [23.84]
+
+[23.84] 好像是正戴着痛苦面具 咬牙切齿地准备纹一个临时文身 [31.34]
+[31.64] 因为欧盟的《数字市场法》规定 谷歌、苹果等守门平台公司 [37.33]
+[37.33] 必须在2024年3月前 开放其软件生态系统 [40.82]
+[41.36] Gurman还说 苹果的财报以及在iOS 17.2 beta版中发现的代码 [47.44]
+[47.44] 都表明苹果正准备在截止日期前 几个月内 允许iPhone侧加载应用 [54.22]
+
+[54.66] 但是 与他们在iPhone上推出USB-C的做法不同 [57.54]
+[58.4] 同样也是因为受到了欧盟的压力 [60.18]
+[60.4] 但侧加载可能只会在欧盟内部开放 [64.1]
+[64.38] 但是虽然苹果准备做出这一惊天壮举 [67.42]
+[67.42] 让他们的用户为自己做决定 [72.22]
+[72.54] 这家万亿美元的公司据报道还在悄悄筹谋绝地反击 提起数项上诉 [79.02]
+[79.26] 挑战欧盟将iOS应用商店和iMessage 定性为守门类服务的决定 [85.3]
+[86.04] 苹果并不认为自己在守门式地管控什么 [89.0]
+
+[89.0] 如果他们真有 也顶多算是霸道总裁的那种溺爱 [92.66]
+[92.98] 就像是哇姐妹你说的好对 [94.6]
+[95.02] 什么建设性建议 都是假的 [96.9]
+[97.36] 总之 [97.64]
+[97.64] 准备为你使用iPhone的朋友们 提供更高级的技术支持吧 [101.44]
+[101.7] 苹果的其他相关新闻 [102.6]
+[102.6] 关于该公司底配M3 MacBook Pro 内存只有8G的担忧 [109.16]
+[109.16] 抱歉 [109.62]
+
+[110.14] 应该是“统一内存” [111.02]
+[111.22] 这些担忧被博主Max Tech的实验 一定程度地证实了 [114.78]
+[115.04] 他让一台8GB内存和一台16GB内存的M3 MacBook Pro进行了对比 [119.0]
+[119.0] 并且发现 毫不意外地 [122.44]
+[122.44] 16G版笔记本的性能几乎吊打8G版 [127.74]
+[127.74] 导出照片和视频的速度提升多达3倍 [131.32]
+[131.32] 尽管它只有前者两倍的容量 [134.16]
+[134.92] 现在 [135.12]
+
+[135.12] 果粉只会对苹果的低能MacBook Pro更加失望 [140.52]
+[140.52] 尤其是在得知该公司的全球市场营销副总裁Bob Borchers [144.66]
+[145.0] 对着Bilibili UP主林亦LYi狺狺狂吠之后 [150.52]
+[151.2] 说什么M3 MacBook Pro 上的8GB RAM [153.96]
+[153.96] 相当于其他系统上的16GB RAM [158.32]
+[158.6] 现在 [159.0]
+[159.0] 你可以承认 苹果的软硬件优化是高科技 [163.82]
+[163.82] 但当你听到Borchers的发言时 你也可以庆幸自己在没喝2%的脱脂牛奶 [168.96]
+[169.18] 我一天内换裤子的次数已经够多了 好吧？ [172.58]
+
+[172.92] 但嘿 这回至少苹果用了两个单独的固态存储模块 [177.62]
+[178.14] 不是因为大家要求他们做的 [179.86]
+[179.86] 他们才不在乎呢 [182.18]
+[182.7] 英特尔一直在等机会 因为一些不知所谓的事情被报道 [185.72]
+[185.72] 所以谢天谢地MSI发布了一个BIOS更新 [188.84]
+[188.84] 使其的Z790系列主板支持 [191.78]
+[191.78] 名字听上去非常盗版的英特尔应用优化功能 简称APO [197.66]
+[197.66] 这个功能实际上挺不错的 [200.14]
+[200.14] Hardware Unboxed在他们更新了的MSI主板上进行了一些测试 [203.44]
+
+[203.44] 并认为APO修复了英特尔E效能核心 处理游戏的方式 [209.64]
+[209.9] 但尽管听上去 这应该适用于几乎所有包含E核心的芯片 [215.06]
+[215.06] 比如12代和13代系列 [217.4]
+[217.72] APO只适用于部分14代酷睿处理器 [221.64]
+[221.92] 当Hardware Unboxed询问英特尔 这是怎么回事时 [225.5]
+[225.92] 英特尔回了一句让人血寒的公司豪言 [228.7]
+[229.42] "英特尔APO没有计划支持更早的产品" [233.88]
+[234.38] 喂？ [234.64]
+
+[234.98] 我是个人啊 [236.28]
+[236.44] 我有情绪的 [237.44]
+[237.7] 嗡嗡 嗡嗡 [238.48]
+[239.0] Hardware Unboxed还问Intel [240.22]
+[240.22] 是否计划支持更多能兼容APO的游戏 现在只有两款 [244.82]
+[245.14] 现在只有《地铁：离去》和《彩虹六号：围攻》支持 [247.08]
+[247.08] 对此 英特尔回答 “有可能” [250.26]
+
+[289.02] 科技速递伙计 他们 很酷的 你懂的 [292.54]
+[292.54] 今天我有点匆忙 [293.34]
+[293.34] 没时间想什么牛X的 [294.86]
+[294.86] 好吧 [295.2]
+[295.2] NVIDIA发布了H200 [297.02]
+[297.02] H100的升级版 [299.0]
+[299.92] 因为当下的AI热潮供不应求 也许美元不久之后就会跟它挂钩 [305.4]
+[305.4] H200 GPU的规格似乎与H100基本相似 [309.46]
+[309.66] 除了显存 [311.08]
+
+[311.08] 从HBM3升级到了HBM3e [314.78]
+[314.78] 提高了显存的带宽 以及频率和容量 [317.56]
+[317.82] H100有80GB的显存 [320.26]
+[320.26] 而H200有141GB [322.98]
+[323.26] 想想看 你可以用它做很多事 [325.44]
+[325.44] 比如模拟8.88GB的M3 MacBook Pro [329.6]
+[329.84] 是的 [330.72]
+[330.72] 你自己算去吧 [331.66]
+[331.66] RTX 4090仍然因为融化电源接头的问题堆满了维修店 [335.96]
+
+[335.96] 在该显卡12V接头的问题被报导一年多后仍旧如此 [341.1]
+[341.44] 根据YouTube频道Northridge Fix的Alex的说法 [344.36]
+[344.36] 他说他的修理店每周仍然会收到大约20到25台RTX 4090 [349.48]
+[349.88] 这一次次地证明 4090的机主们 [352.36]
+[352.36] 只是没有正确地插上电源而已 你只需用力推到底 [356.08]
+[356.08] 那是你的锅 [356.78]
+[357.0] 值得庆幸的是 [357.42]
+
+[357.42] PCI SIG现已对接头进行了重新设计 [360.14]
+[360.14] 所以希望未来犹豫着的割手党 [364.18]
+[364.18] 不用承受你们现在的痛苦 [366.18]
+[366.4] 其实那并不是你的锅 [367.94]
+[367.94] 我只是开个玩笑 [368.56]
+[368.56] 别写毒评论了 [369.78]
+[370.14] 三天前的网络攻击已经结束 澳大利亚的港口重新开放 [375.68]
+[375.86] 根据DP World Australia的说法 [377.78]
+[377.98] 跟你刚想的那个不沾边 [380.62]
+[380.62] 它实际上是澳大利亚最大的港口运营商 [383.34]
+
+[383.66] 听起来还是挺… [386.04]
+[386.04] 淦 [386.5]
+[387.0] 此次网络攻击导致严重拥堵 且一些数据据称已被攻击者窃取 [391.86]
+[391.86] 攻击者被认为是受境外势力指使 [394.36]
+[394.68] 但我想他们应该看看 船头是不是炸了 [397.58]
+[397.58] 故障频出的闪迪 Extreme Pro SSD [400.99]
+[400.99] 是无法通过任何固件更新修复的硬件问题 [405.32]
+[405.52] 根据数据恢复公司Atingo的说法 [408.36]
+[408.36] 该公司的总经理告诉Futurezone 一些元件 [411.9]
+[411.9] 对于电路板来说 太大了 [415.1]
+
+[415.1] 这导致连接不稳定 [416.58]
+[417.0] 从而造成一些元件完全从电路板上掉了下来 [420.16]
+[420.38] 而且不是以好的方式 [421.68]
+[421.68] 我肯定会十分惊讶 直到我记起来Sandisk是WD的子公司 [427.48]
+[427.48] 而特斯拉Cybertruck的订购协议禁止车主 [431.81]
+[431.81] 在交付后一年内出售这辆多边形车 [436.64]
+[437.04] 除非你告诉特斯拉你想卖 并给他们回购的机会 [441.58]
+[441.58] 如果他们不想回购的话 [443.56]
+[443.56] 他们可能会给你书面许可 [445.52]
+[445.52] 转手卖给第三方 [446.86]
+
+[447.12] 像是什么特许执照 [447.86]
+[448.02] 但为的是应得的基本经济权利 [449.72]
+[449.72] 如果违反了协议的任何条款 [453.12]
+[453.48] 特斯拉称他们可以通过定位并远程停用车辆 [457.28]
+[457.7] 而且说实话 如果你购买了Cybertruck 这些听起来都很说得过去 [461.16]
+[461.52] 它不是一辆普通的皮卡 [462.62]
+[463.14] 它是... [464.02]
+
+[464.02] 它...很丑 [465.46]
+[465.46] 但是你内外都很美 你知道的 所以我知道你会在周三回来看科技新闻的 [470.26]
+[470.26] 这是你的风格 你懂吗 我们喜欢你这样 [473.24]
+[473.24] 拜托了 [473.7]

data/2023-11-13 The EU Will Break Apple. [Hod69Y1aVlY].en.txt

@@ -0,0 +1,172 @@
+[0.0] Uh, [0.48]
+[0.48] you came here because you want tech news, [3.76]
+[3.96] and I know that. [5.4]
+[5.92] And that's why I have that ready for, [9.16]
+[9.16] I have it ready for you right here. [10.66]
+[10.66] It's just, one second. [12.42]
+[12.84] Apple doesn't  want to allow iPhone users [16.06]
+[16.06] to install apps from outside of the app store, [18.12]
+[18.12] AKA side loading, [19.54]
+[19.54] but according to Mark Gurm's together strong Gurman, [23.06]
+[23.06] the company looks [23.84]
+
+[23.84] like it's doing the tech giant equivalent [25.64]
+[25.64] of gritting its teeth and wincing in anticipation [28.62]
+[29.0] of getting a very painful temporary tattoo. [31.34]
+[31.64] See, the EU's Digital Market Act requires gatekeeper companies like Google and Apple [37.33]
+[37.33] to open up their software ecosystems by March, 2024. [40.82]
+[41.36] The Gurman, financial filings from Apple, [43.9]
+[43.9] as well as code found in the iOS 17.2 beta, [47.44]
+[47.44] all point to Apple preparing to allow side loading on iPhones before that deadline, [52.8]
+[52.8] so in the next few months. [54.22]
+
+[54.66] However, unlike their rollout of USB-C on iPhones, [57.54]
+[58.4] which happened because of EU pressure, [60.18]
+[60.4] side loading may only be allowed in the EU itself. [64.1]
+[64.38] But while Apple is preparing to do the unthinkable, [67.42]
+[67.42] letting their users decide something for  themselves, [72.22]
+[72.54] the trillion dollar company is reportedly also quietly cooking up a Hail Mary in the form of several appeals, [79.02]
+[79.26] challenging the EU's decision to label the iOS app store and iMessage as gatekeeper services. [85.3]
+[86.04] Apple doesn't think of itself as gatekeeping anything. [89.0]
+
+[89.0] And if they are, it's more like a girl boss gaslight situation. [92.66]
+[92.98] Like, wow, yes, queen. [94.6]
+[95.02] Constructive criticism isn't real. [96.9]
+[97.36] Anyway, [97.64]
+[97.64] get ready to provide a whole new level of tech support for your iPhone using friends. [101.44]
+[101.7] In other Apple news, [102.6]
+[102.6] concerns over the company's decision [104.16]
+[104.16] to limit the base model M3 MacBook Pro to eight gigabytes of RAM, [109.16]
+[109.16] sorry, [109.62]
+
+[110.14] unified memory, [111.02]
+[111.22] have been somewhat validated, [112.8]
+[112.8] according to YouTuber Max Tech. [114.78]
+[115.04] He put an eight gigabyte and a 16 gigabyte M3 MacBook Pro head to head, [119.0]
+[119.0] and found that, [119.96]
+[119.96] unsurprising to nearly everyone, [122.44]
+[122.44] the 16 gigabyte laptop enjoys a sizable performance boost over the eight gigabyte model, [127.74]
+[127.74] exporting photos and videos as much as three times faster, [131.32]
+[131.32] despite only having two times the remembering power. [134.16]
+[134.92] Now, [135.12]
+
+[135.12] enthusiasts only got more upset about Apple's decision to under-equip the base model MacBook Pro [140.52]
+[140.52] after the company's VP [141.96]
+[141.96] of worldwide marketing, Bob Borchers, [144.66]
+[145.0] totally borked himself by telling bilibili creator,  Yee Lee [150.52]
+[151.2] that eight gigabytes of RAM on the M3 MacBook [153.96]
+[153.96] Pro is analogous to 16 gigabytes on other systems. [158.32]
+[158.6] Now, [159.0]
+[159.0] you can acknowledge that Apple's software and hardware optimization is high art, [163.82]
+[163.82] while also being glad you weren't drinking some delicious 2% milk when you heard Borchers claim. [168.96]
+[169.18] Because I changed my pants enough times in one day, okay? [172.58]
+
+[172.92] But hey, at least Apple used two separate NAND modules for the storage this time. [177.62]
+[178.14] Not because people asked for that. [179.86]
+[179.86] They don't care. [182.18]
+[182.7] Intel has been waiting its turn to be called out for some nonsense, [185.72]
+[185.72] so thank goodness that MSI released a BIOS update [188.84]
+[188.84] for its Z790 motherboards that enables support [191.78]
+[191.78] for the incredibly generic sounding Intel Application Optimization feature, or APO. [197.66]
+[197.66] The feature is actually pretty cool. [200.14]
+[200.14] Hardware Unboxed did some testing with their now updated MSI boards, [203.44]
+
+[203.44] and credited APO as a fix for the way Intel's efficiency cores, or E-cores, [208.84]
+[209.16] handle gaming. [209.64]
+[209.9] But despite sounding like something that should work on pretty much any Intel chip with efficiency cores, [215.06]
+[215.06] like the 12th and 13th gen lineups, [217.4]
+[217.72] APO only works on select 14th gen core processors. [221.64]
+[221.92] And when Hub asked Intel, [223.98]
+[224.18] what's the deal with that? [225.5]
+[225.92] Intel gave a bone chillingly corporate response. [228.7]
+[229.42] Intel has no plans to support prior generation products with application optimization. [233.88]
+[234.38] Hello? [234.64]
+
+[234.98] I'm a human being. [236.28]
+[236.44] I have feelings. [237.44]
+[237.7] Beep boop. [238.48]
+[239.0] Hardware Unboxed also asked Intel [240.22]
+[240.22] if they plan to support more than the only two games that work with APO right now, [244.82]
+[245.14] Metro Exodus and Rainbow Six Siege, [247.08]
+[247.08] to which Intel replied, yeah, probably. [250.26]
+
+[289.02] Quick Bits, man, they are, they're so cool. [292.54]
+[292.54] I'm in a rush today. [293.34]
+[293.34] I didn't have time to think of something amazing. [294.86]
+[294.86] Okay, well, [295.2]
+[295.2] NVIDIA has announced the H200, [297.02]
+[297.02] an upgraded version of the H100 GPU. [299.0]
+[299.92] which is so in demand for running AI right now that the US's dollar value [304.02]
+[304.02] might get tethered to it soon. [305.4]
+[305.4] The H200's specs appear to be largely similar to the H100's, [309.46]
+[309.66] except for the memory, [311.08]
+
+[311.08] which is upgraded from HBM3 to HBM3e, [314.78]
+[314.78] boosting memory bandwidth, clocks, and capacity. [317.56]
+[317.82] The H100 had 80 gigabytes of VRAM, [320.26]
+[320.26] while the H200 has 141. [322.98]
+[323.26] Think about all the things you could do with that. [325.44]
+[325.44] Like emulate 8.88 gigabyte M3 Macbook Pros [329.6]
+[329.84] Yeah, [330.72]
+[330.72] you're gonna do the math. [331.66]
+[331.66] RTX 4090s are still pouring into repair shops  with melted power plugs [335.96]
+
+[335.96] more than a year after problems with the card's 12 volt high power connector were first reported, [341.1]
+[341.44] according to Alex from Northridge Fix on YouTube. [344.36]
+[344.36] He says his repair shop still gets [346.06]
+[346.06] about 20 to 25 4090s in per week, [349.48]
+[349.88] proving once and for all that the 4090 owners [352.36]
+[352.36] just didn't plug them in right, guys. [354.32]
+[354.32] You gotta push it all the way in. [356.08]
+[356.08] It's your fault. [356.78]
+[357.0] Thankfully, [357.42]
+
+[357.42] PCI SIG has now redesigned the connector, [360.14]
+[360.14] so hopefully future generations of financially questionable decision makers [364.18]
+[364.18] won't endure the suffering you have. [366.18]
+[366.4] It's not actually your fault. [367.94]
+[367.94] I was joking. [368.56]
+[368.56] Stop writing that comment. [369.78]
+[370.14] Australia's ports are reopening after being shut down by a massive cyber attack three days ago, [375.68]
+[375.86] according to DP World Australia, [377.78]
+[377.98] which is not related to what you're thinking of. [380.62]
+[380.62] It's actually the country's biggest ports operator, [383.34]
+
+[383.66] which still sounds... [386.04]
+[386.04] Damn it. [386.5]
+[387.0] The cyber attack caused huge delays and congestion, [389.2]
+[389.36] and some data was apparently stolen by the perpetrators, [391.86]
+[391.86] who are believed to be foreign state actors. [394.36]
+[394.68] But I reckon they should see whether the front fell off. [397.58]
+[397.58] Widely reported issues with SanDisk Extreme Pro SSDs [400.99]
+[400.99] are the result of hardware issues unfixable by any firmware update, [405.32]
+[405.52] according to data recovery company Atingo. [408.36]
+[408.36] Their managing director told Futurezone that some of the components [411.9]
+[411.9] SanDisk used are too big for their circuit boards, [415.1]
+
+[415.1] leading to unstable connections, [416.58]
+[417.0] which caused some components to pop off the board entirely, [420.16]
+[420.38] and not in a good way. [421.68]
+[421.68] I would be shocked and appalled by this if I didn't remember that SanDisk is owned by Western Digital. [427.48]
+[427.48] And Tesla's order agreement for the Cybertruck forbids owners [431.81]
+[431.81] from selling the polygonal vehicle for one year after it's delivered, [436.64]
+[437.04] unless you tell Tesla you want to sell it and give them a chance to buy it back from you. [441.58]
+[441.58] If they don't want to buy it back, [443.56]
+[443.56] they might give you written permission [445.52]
+[445.52] to sell it to a third party, [446.86]
+
+[447.12] like a hall pass, [447.86]
+[448.02] but for basic economic rights. [449.72]
+[449.72] If you violate this or any other provision in the agreement, [453.12]
+[453.48] Tesla says they may locate and disable the vehicle electronically. [457.28]
+[457.7] And really, I mean, [458.84]
+[458.84] this should all make sense if you bought a Cybertruck. [461.16]
+[461.52] It's not a normal truck. [462.62]
+[463.14] It's... [464.02]
+
+[464.02] It's ugly. [465.46]
+[465.46] But you're beautiful inside and out, [467.48]
+[467.48] and that's why I know you're gonna come back on Wednesday for more tech news. [470.26]
+[470.26] It's just who you are, you know. [471.64]
+[471.64] We love you for that. [473.24]
+[473.24] Please. [473.7]

generate_chatgpt.py

@@ -0,0 +1,126 @@
+import json
+import os
+from os import listdir
+from os.path import isfile, join
+
+INSTRUCT_CHUNKED_PROMPT = """你是一个擅长翻译科技新闻的翻译专家。请将以下内容翻译为中文，使用相同格式输出，并保留时间戳。不要漏掉任何信息。合并多行文本时，保留第一个和最后一个时间戳。
+"""
+
+def new_message(eng_in, chs_out, prev_in = None, prev_out = None):
+    if(prev_in == None or prev_out == None):
+        return {"messages": [
+            {"role": "system", "content": INSTRUCT_CHUNKED_PROMPT}, 
+            {"role": "user", "content": eng_in}, 
+            {"role": "assistant", "content": chs_out}]
+        }
+    else:
+        return {"messages": [
+            {"role": "system", "content": INSTRUCT_CHUNKED_PROMPT}, 
+            {"role": "user", "content": prev_in}, 
+            {"role": "assistant", "content": prev_out},
+            {"role": "user", "content": eng_in}, 
+            {"role": "assistant", "content": chs_out}]
+        }
+
+def write_jsonl(message_groups, filename):
+    with open(filename, "w", encoding='utf-8-sig') as fout:
+        for i in range(len(message_groups)):
+            if(i>0):
+                msg_obj = new_message(
+                    message_groups[i][0].strip(),
+                    message_groups[i][1].strip(),
+                    message_groups[i-1][0].strip(),
+                    message_groups[i-1][1].strip()
+                )
+            else:
+                msg_obj = new_message(
+                    message_groups[i][0].strip(),
+                    message_groups[i][1].strip()
+                )
+            json.dump(msg_obj, fout)
+            fout.write("\n")
+
+message_groups = []
+
+DOCUMENT_ROOT = "data"
+files = listdir("data")
+files = list(filter(lambda x: x.endswith(".en.txt"), files))
+files.sort()
+
+print(files)
+
+for f in files:
+    en_fname = join(DOCUMENT_ROOT, f)
+    if en_fname.endswith(".en.txt") and isfile(en_fname):
+        cn_fname = join(DOCUMENT_ROOT, f.replace(".en.txt", ".cn.txt"))
+        if os.path.exists(cn_fname) and isfile(cn_fname):
+            print(f"Found data pair: {en_fname} and {cn_fname}")
+
+            with open(en_fname, "r", encoding='utf-8-sig') as enfin:
+                en_messages = enfin.read()
+
+            with open(cn_fname, "r", encoding='utf-8-sig') as cnfin:
+                cn_messages = cnfin.read()
+            
+            en_messages = [part.strip() for part in en_messages.split("\n\n") if part.strip() != ""]
+            cn_messages = [part.strip() for part in cn_messages.split("\n\n") if part.strip() != ""]
+
+            if(len(en_messages) != len(cn_messages)):
+                print(f"English and Chinese version mismatch. Discarding {en_fname} pair.")
+            
+            messages = zip(en_messages, cn_messages)
+
+            message_groups.extend(messages)
+
+write_jsonl(message_groups, "combined.jsonl")
+
+import random
+random.shuffle(message_groups)
+
+TEST_RATIO = 0.2
+
+split_index = int(len(message_groups) * TEST_RATIO)
+
+test = message_groups[:split_index]
+train = message_groups[split_index:]
+
+write_jsonl(train, "chatgpt-train.jsonl")
+write_jsonl(test, "chatgpt-test.jsonl")
+
+recent_files = files[-5:]
+recent_messages = []
+
+for f in recent_files:
+    en_fname = join(DOCUMENT_ROOT, f)
+    if en_fname.endswith(".en.txt") and isfile(en_fname):
+        cn_fname = join(DOCUMENT_ROOT, f.replace(".en.txt", ".cn.txt"))
+        if os.path.exists(cn_fname) and isfile(cn_fname):
+            print(f"Found data pair: {en_fname} and {cn_fname}")
+
+            with open(en_fname, "r", encoding='utf-8-sig') as enfin:
+                en_messages = enfin.read()
+
+            with open(cn_fname, "r", encoding='utf-8-sig') as cnfin:
+                cn_messages = cnfin.read()
+            
+            en_messages = [part.strip() for part in en_messages.split("\n\n") if part.strip() != ""]
+            cn_messages = [part.strip() for part in cn_messages.split("\n\n") if part.strip() != ""]
+
+            if(len(en_messages) != len(cn_messages)):
+                print(f"English and Chinese version mismatch. Discarding {en_fname} pair.")
+            
+            messages = zip(en_messages, cn_messages)
+
+            recent_messages.extend(messages)
+
+write_jsonl(recent_messages, "recent-combined.jsonl")
+
+TEST_RATIO = 0.2
+
+split_index = int(len(recent_messages) * TEST_RATIO)
+
+test = recent_messages[:split_index]
+train = recent_messages[split_index:]
+
+write_jsonl(train, "chatgpt-recent-train.jsonl")
+write_jsonl(test, "chatgpt-recent-test.jsonl")

generate_llamafactory.py

@@ -0,0 +1,60 @@
+import os
+from os import listdir
+from os.path import isfile, join
+import json
+
+INSTRUCT_CHUNKED_PROMPT = """你是一个擅长翻译科技新闻的翻译专家。请将以下内容翻译为中文，使用相同格式输出，并保留时间戳。不要漏掉任何信息。合并多行文本时，保留第一个和最后一个时间戳。"""
+
+DOCUMENT_ROOT = "data"
+files = listdir("data")
+
+all_datapoints = []
+
+for f in files:
+    en_fname = join(DOCUMENT_ROOT, f)
+    if en_fname.endswith(".en.txt") and isfile(en_fname):
+        cn_fname = join(DOCUMENT_ROOT, f.replace(".en.txt", ".cn.txt"))
+        if os.path.exists(cn_fname) and isfile(cn_fname):
+            print(f"Found data pair: {en_fname} and {cn_fname}")
+
+            with open(en_fname, "r", encoding='utf-8-sig') as enfin:
+                en_messages = enfin.read()
+
+            with open(cn_fname, "r", encoding='utf-8-sig') as cnfin:
+                cn_messages = cnfin.read()
+            
+            en_messages = [part.strip() for part in en_messages.split("\n\n") if part.strip() != ""]
+            cn_messages = [part.strip() for part in cn_messages.split("\n\n") if part.strip() != ""]
+
+            if(len(en_messages) != len(cn_messages)):
+                print(f"English and Chinese version mismatch. Discarding {en_fname} pair.")
+            
+            # Write document
+            history = []
+            for en, cn in zip(en_messages, cn_messages):
+                all_datapoints.append({
+                    'system': INSTRUCT_CHUNKED_PROMPT, 
+                    'input': en,
+                    'output': cn,
+                    'history': history
+                })
+                history.append([en, cn])
+
+import random
+shuffled = random.shuffle(all_datapoints)
+
+COUNT = len(all_datapoints)
+TRAIN = int(0.8 * COUNT)
+
+train_dataset = all_datapoints[:TRAIN]
+evaluation_dataset = all_datapoints[TRAIN:]
+
+with open("train.json", "w", encoding='utf-8-sig') as train_f:
+    for d in train_dataset:
+        json.dump(d, train_f)
+        train_f.write('\n')
+
+with open("test.json", "w", encoding='utf-8-sig') as ev_f:
+    for d in evaluation_dataset:
+        json.dump(d, ev_f)
+        ev_f.write('\n')

jsonl_gen.ipynb

@@ -1,229 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import json"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 10,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "INSTRUCT_CHUNKED_PROMPT = \"\"\"你是一个擅长翻译科技新闻的翻译专家。请将以下内容翻译为中文，使用相同格式输出，并保留时间戳。不要漏掉任何信息。合并多行文本时，保留第一个和最后一个时间戳。\n",
-    "\"\"\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 11,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "def new_message(eng_in, chs_out, prev_in = None, prev_out = None):\n",
-    "    if(prev_in == None or prev_out == None):\n",
-    "        return {\"messages\": [\n",
-    "            {\"role\": \"system\", \"content\": INSTRUCT_CHUNKED_PROMPT}, \n",
-    "            {\"role\": \"user\", \"content\": eng_in}, \n",
-    "            {\"role\": \"assistant\", \"content\": chs_out}]\n",
-    "        }\n",
-    "    else:\n",
-    "        return {\"messages\": [\n",
-    "            {\"role\": \"system\", \"content\": INSTRUCT_CHUNKED_PROMPT}, \n",
-    "            {\"role\": \"user\", \"content\": prev_in}, \n",
-    "            {\"role\": \"assistant\", \"content\": prev_out},\n",
-    "            {\"role\": \"user\", \"content\": eng_in}, \n",
-    "            {\"role\": \"assistant\", \"content\": chs_out}]\n",
-    "        }"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 12,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "def write_jsonl(message_groups, filename):\n",
-    "    with open(filename, \"w\") as fout:\n",
-    "        for i in range(len(message_groups)):\n",
-    "            if(i>0):\n",
-    "                msg_obj = new_message(\n",
-    "                    message_groups[i][0].strip(),\n",
-    "                    message_groups[i][1].strip(),\n",
-    "                    message_groups[i-1][0].strip(),\n",
-    "                    message_groups[i-1][1].strip()\n",
-    "                )\n",
-    "            else:\n",
-    "                msg_obj = new_message(\n",
-    "                    message_groups[i][0].strip(),\n",
-    "                    message_groups[i][1].strip()\n",
-    "                )\n",
-    "            json.dump(msg_obj, fout)\n",
-    "            fout.write(\"\\n\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 13,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "message_groups = []"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 14,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "['2023-10-11 Watch Out, Apple... [5v72vj2nFN0].en.txt', \"2023-10-14 It's Finally Over. [KJF5Nn23CwM]-4k.en.txt\", '2023-10-16 Apple, Are U Ok？ [34R2xVzBmfA].en.txt', '2023-10-18 China lost its GPU privileges [Wz9JoTTRa4k].en.txt', \"2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].en.txt\", '2023-10-28 Google Pays Apple EVERY YEAR [qvZAYJLmzuk].en.txt', '2023-10-30 Nvidia Please No [AQ8kEsu41uE].en.txt', '2023-11-01 Gamers： Apple Wants YOU! [BIQ3JcWJzT0].en.txt', '2023-11-04 Actually, Google LIKES Adblock [Z6zZFj9eE60].en.txt', '2023-11-06 Grok [rCmsoa2Zlhg].en.txt']\n",
-      "Found data pair: data\\2023-10-11 Watch Out, Apple... [5v72vj2nFN0].en.txt and data\\2023-10-11 Watch Out, Apple... [5v72vj2nFN0].cn.txt\n",
-      "Found data pair: data\\2023-10-14 It's Finally Over. [KJF5Nn23CwM]-4k.en.txt and data\\2023-10-14 It's Finally Over. [KJF5Nn23CwM]-4k.cn.txt\n",
-      "Found data pair: data\\2023-10-16 Apple, Are U Ok？ [34R2xVzBmfA].en.txt and data\\2023-10-16 Apple, Are U Ok？ [34R2xVzBmfA].cn.txt\n",
-      "Found data pair: data\\2023-10-18 China lost its GPU privileges [Wz9JoTTRa4k].en.txt and data\\2023-10-18 China lost its GPU privileges [Wz9JoTTRa4k].cn.txt\n",
-      "Found data pair: data\\2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].en.txt and data\\2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].cn.txt\n",
-      "Found data pair: data\\2023-10-28 Google Pays Apple EVERY YEAR [qvZAYJLmzuk].en.txt and data\\2023-10-28 Google Pays Apple EVERY YEAR [qvZAYJLmzuk].cn.txt\n",
-      "Found data pair: data\\2023-10-30 Nvidia Please No [AQ8kEsu41uE].en.txt and data\\2023-10-30 Nvidia Please No [AQ8kEsu41uE].cn.txt\n",
-      "Found data pair: data\\2023-11-01 Gamers： Apple Wants YOU! [BIQ3JcWJzT0].en.txt and data\\2023-11-01 Gamers： Apple Wants YOU! [BIQ3JcWJzT0].cn.txt\n",
-      "Found data pair: data\\2023-11-04 Actually, Google LIKES Adblock [Z6zZFj9eE60].en.txt and data\\2023-11-04 Actually, Google LIKES Adblock [Z6zZFj9eE60].cn.txt\n",
-      "Found data pair: data\\2023-11-06 Grok [rCmsoa2Zlhg].en.txt and data\\2023-11-06 Grok [rCmsoa2Zlhg].cn.txt\n"
-     ]
-    }
-   ],
-   "source": [
-    "import os\n",
-    "from os import listdir\n",
-    "from os.path import isfile, join\n",
-    "\n",
-    "DOCUMENT_ROOT = \"data\"\n",
-    "files = listdir(\"data\")\n",
-    "files = list(filter(lambda x: x.endswith(\".en.txt\"), files))\n",
-    "files.sort()\n",
-    "\n",
-    "print(files)\n",
-    "\n",
-    "for f in files:\n",
-    "    en_fname = join(DOCUMENT_ROOT, f)\n",
-    "    if en_fname.endswith(\".en.txt\") and isfile(en_fname):\n",
-    "        cn_fname = join(DOCUMENT_ROOT, f.replace(\".en.txt\", \".cn.txt\"))\n",
-    "        if os.path.exists(cn_fname) and isfile(cn_fname):\n",
-    "            print(f\"Found data pair: {en_fname} and {cn_fname}\")\n",
-    "\n",
-    "            with open(en_fname, \"r\") as enfin:\n",
-    "                en_messages = enfin.read()\n",
-    "\n",
-    "            with open(cn_fname, \"r\") as cnfin:\n",
-    "                cn_messages = cnfin.read()\n",
-    "            \n",
-    "            en_messages = [part.strip() for part in en_messages.split(\"\\n\\n\") if part.strip() != \"\"]\n",
-    "            cn_messages = [part.strip() for part in cn_messages.split(\"\\n\\n\") if part.strip() != \"\"]\n",
-    "\n",
-    "            if(len(en_messages) != len(cn_messages)):\n",
-    "                print(f\"English and Chinese version mismatch. Discarding {en_fname} pair.\")\n",
-    "            \n",
-    "            messages = zip(en_messages, cn_messages)\n",
-    "\n",
-    "            message_groups.extend(messages)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 15,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "write_jsonl(message_groups, \"combined.jsonl\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 16,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import random\n",
-    "random.shuffle(message_groups)\n",
-    "\n",
-    "TEST_RATIO = 0.2\n",
-    "\n",
-    "split_index = int(len(message_groups) * TEST_RATIO)\n",
-    "\n",
-    "test = message_groups[:split_index]\n",
-    "train = message_groups[split_index:]\n",
-    "\n",
-    "write_jsonl(train, \"chatgpt-train.jsonl\")\n",
-    "write_jsonl(test, \"chatgpt-test.jsonl\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 20,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "recent_files = files[-5:]\n",
-    "recent_messages = []"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "for f in recent_files:\n",
-    "    en_fname = join(DOCUMENT_ROOT, f)\n",
-    "    if en_fname.endswith(\".en.txt\") and isfile(en_fname):\n",
-    "        cn_fname = join(DOCUMENT_ROOT, f.replace(\".en.txt\", \".cn.txt\"))\n",
-    "        if os.path.exists(cn_fname) and isfile(cn_fname):\n",
-    "            print(f\"Found data pair: {en_fname} and {cn_fname}\")\n",
-    "\n",
-    "            with open(en_fname, \"r\") as enfin:\n",
-    "                en_messages = enfin.read()\n",
-    "\n",
-    "            with open(cn_fname, \"r\") as cnfin:\n",
-    "                cn_messages = cnfin.read()\n",
-    "            \n",
-    "            en_messages = [part.strip() for part in en_messages.split(\"\\n\\n\") if part.strip() != \"\"]\n",
-    "            cn_messages = [part.strip() for part in cn_messages.split(\"\\n\\n\") if part.strip() != \"\"]\n",
-    "\n",
-    "            if(len(en_messages) != len(cn_messages)):\n",
-    "                print(f\"English and Chinese version mismatch. Discarding {en_fname} pair.\")\n",
-    "            \n",
-    "            messages = zip(en_messages, cn_messages)\n",
-    "\n",
-    "            recent_messages.extend(messages)"
-   ]
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "whisper",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.10.13"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}

jsonl_gen_for_chatglm.ipynb

@@ -1,120 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Found data pair: data\\2023-10-11 Watch Out, Apple... [5v72vj2nFN0].en.txt and data\\2023-10-11 Watch Out, Apple... [5v72vj2nFN0].cn.txt\n",
-      "Found data pair: data\\2023-10-14 It's Finally Over. [KJF5Nn23CwM]-4k.en.txt and data\\2023-10-14 It's Finally Over. [KJF5Nn23CwM]-4k.cn.txt\n",
-      "Found data pair: data\\2023-10-16 Apple, Are U Ok？ [34R2xVzBmfA].en.txt and data\\2023-10-16 Apple, Are U Ok？ [34R2xVzBmfA].cn.txt\n",
-      "Found data pair: data\\2023-10-18 China lost its GPU privileges [Wz9JoTTRa4k].en.txt and data\\2023-10-18 China lost its GPU privileges [Wz9JoTTRa4k].cn.txt\n",
-      "Found data pair: data\\2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].en.txt and data\\2023-10-25 Windows' Apple Silicon Moment is Coming [xFvGERpRUpM].cn.txt\n",
-      "Found data pair: data\\2023-10-28 Google Pays Apple EVERY YEAR [qvZAYJLmzuk].en.txt and data\\2023-10-28 Google Pays Apple EVERY YEAR [qvZAYJLmzuk].cn.txt\n"
-     ]
-    }
-   ],
-   "source": [
-    "import os\n",
-    "from os import listdir\n",
-    "from os.path import isfile, join\n",
-    "import json\n",
-    "\n",
-    "INSTRUCT_CHUNKED_PROMPT = \"\"\"你是一个擅长翻译科技新闻的翻译专家。请将以下内容翻译为中文，使用相同格式输出，并保留时间戳。不要漏掉任何信息。合并多行文本时，保留第一个和最后一个时间戳。\"\"\"\n",
-    "\n",
-    "DOCUMENT_ROOT = \"data\"\n",
-    "files = listdir(\"data\")\n",
-    "\n",
-    "all_datapoints = []\n",
-    "\n",
-    "for f in files:\n",
-    "    en_fname = join(DOCUMENT_ROOT, f)\n",
-    "    if en_fname.endswith(\".en.txt\") and isfile(en_fname):\n",
-    "        cn_fname = join(DOCUMENT_ROOT, f.replace(\".en.txt\", \".cn.txt\"))\n",
-    "        if os.path.exists(cn_fname) and isfile(cn_fname):\n",
-    "            print(f\"Found data pair: {en_fname} and {cn_fname}\")\n",
-    "\n",
-    "            with open(en_fname, \"r\") as enfin:\n",
-    "                en_messages = enfin.read()\n",
-    "\n",
-    "            with open(cn_fname, \"r\") as cnfin:\n",
-    "                cn_messages = cnfin.read()\n",
-    "            \n",
-    "            en_messages = [part.strip() for part in en_messages.split(\"\\n\\n\") if part.strip() != \"\"]\n",
-    "            cn_messages = [part.strip() for part in cn_messages.split(\"\\n\\n\") if part.strip() != \"\"]\n",
-    "\n",
-    "            if(len(en_messages) != len(cn_messages)):\n",
-    "                print(f\"English and Chinese version mismatch. Discarding {en_fname} pair.\")\n",
-    "            \n",
-    "            # Write document\n",
-    "            history = []\n",
-    "            for en, cn in zip(en_messages, cn_messages):\n",
-    "                all_datapoints.append({\n",
-    "                    'system': INSTRUCT_CHUNKED_PROMPT, \n",
-    "                    'input': en,\n",
-    "                    'output': cn,\n",
-    "                    'history': history\n",
-    "                })\n",
-    "                history.append([en, cn])"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import random\n",
-    "shuffled = random.shuffle(all_datapoints)\n",
-    "\n",
-    "COUNT = len(all_datapoints)\n",
-    "TRAIN = int(0.8 * COUNT)\n",
-    "\n",
-    "train_dataset = all_datapoints[:TRAIN]\n",
-    "evaluation_dataset = all_datapoints[TRAIN:]\n",
-    "\n",
-    "with open(\"train.json\", \"w\") as train_f:\n",
-    "    for d in train_dataset:\n",
-    "        json.dump(d, train_f)\n",
-    "        train_f.write('\\n')\n",
-    "\n",
-    "with open(\"test.json\", \"w\") as ev_f:\n",
-    "    for d in evaluation_dataset:\n",
-    "        json.dump(d, ev_f)\n",
-    "        ev_f.write('\\n')\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.9.18"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}

test.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f85f753eeb7220638ac9a806f5b81bb422f2f505ff6f1acac2385c2fc32de060
-size 1341274
+oid sha256:8d5209c74ca98036583b7a7e826e038acc78e32d46d1db526fb71b100b32a2bc
+size 1426136

train.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:57be287bae0ca27540a71b5b1ed1d8b35ac416c72d92a6a75c134cbc4a870eba
-size 5322699
+oid sha256:3b11a55096b154473c92e453a167e015c2af7624845b0b781d4b5c9877c36eb0
+size 5662033