Meta tags:
keywords=Python weblog blog blogs blogger weblogger aggregator rss;
description=Recent postings from Python-related blogs.;

Headings (most frequently used words):

the, and, python, code, for, block, of, you, 2026, in, words, with, about, to, is, bag, on, may, keynote, pycon, what, pytorch, tensorflow, pycharm, can, real, us, any, model, test, at, data, pyconus, how, pablo, textual, or, news, your, project, arrow, asking, they, are, an, conference, open, get, buttons, have, strengths, weaknesses, word, quiz, why, podcast, pep, using, vs, experience, started, this, blog, where, our, speakers, tech, tips, cogs, programming, changes, new, creating, streamlit, content, button, english, work, 15, sprints, first, it, do, that, cursor, batch_size, 8192, pyarrow, tell, think, slow, talking, spy, behind, commit, releases, fourth, final, head, different, use, applications, support, choosing, right, framework, apache, should, series, we, re, each, their, journey, into, excited, provide, awesome, galindo, 16, dialog, examples, natural, language, user, dynamic, based, april, techniques, applying, alternatives, alpha, before, 14, accepted, packaging, council, up, __getattr__, tense, turkish, tryton, handling, europython, vocabulary, stop, lemmatization, tf, idf, ignores, know, now, key, questions, speaker, did, pair, planet, 04, highlights, read, full, article, https, realpython, com, one, deals, others, deal, accessing, dot, notation, horseman, __get__, appendix, blocks, sets, apart, comparison, cases, tooling, developer, performance, scalability, deployment, community, ecosystem, library, fetch, apis, testing, getting, next, note, thanks, resources, rise, llm, slop, engage, farm, speed, kill, friction, rate, limiting, trust, erosion, gaslighting, suggestions, change, 03, thank, interview, learn, more, page, also, attend, salgados, meet, greet, psf, booth, expo, hall, saturday, after, 02, 01, directory, file, install, links, modules, documentation, system, administrator, implementers, developers, multiple, actions, displaying, clicks, forms, press, conclusion, 30, playing, translations, references, gameplay, factions, story, 29, does, advantages, nlp, implementing, visualizing, results, advanced, world, problem, limitations, today, beta, freeze, incremental, gc, reverted, 772, gets, 803, stable, abi, goes, free, threaded, who, which, join, pr, cpython, there, easy, issues, speak, fun, coding, stack, __getattribute__, asymmetry, between, __setattr__, instance, testingattributeaccess, second, third, fifth, daniel, roy, greenfeld, processing, computer, vision, reinforcement, learning, engineering, microsoft, try, share, feedback, armin, ronacher, michiel, rodrigo, girão, serrão, mike, driscoll, luke, plant, temporary, permanent, present, french, gendered, pronouns, nouns, mış, comments, accounting, invoicing, payments, stock, production, shipments, interface, guis, basic, syntax, form, submission, antonio, cuni, seth, michael, larson, tokenization, creation, encoding, setting, loading, preparing, frequency, plots, removal, stemming, grams, out, dimensionality, reduction, feature, selection, architecture, training, validation, generating, predictions, running, baseline, removing, top, terms, vectorization, fitting, revised, set, loses, order, information, semantics, context, result, large, sparse, vectors, other, everything, always, wanted, management, sqlite, sqlalchemy, counter, markdown, introducing, mssql, arrow_batch, recordbatch, table, arrow_reader, recordbatchreader, sake, talk, pygrunn, httpxyz, salgado, revisit, fundamentals, amanda, casari, without, giving, too, many, spoilers, friend, mentor, helped, most, important, ve, ever, done, if, might, still, be, future, something, plans, been, looking, forward, advice, time, goers, general, source, culture, not, enough, people, til, 144, sentinel, built, released, episode, 293, agentic, science, marimo, factory, method, pattern, its, implementation, inverse, sapir, whorf, languages, making, things, happen, analysis, ticket, sales, frog, whom, bell, tolls, call, volunteers, 10, 11, 12, 13, 17, 18, 19, 20, json, serialization, bonus,

Text of the page (most frequently used words):
the (1290), and (605), you (422), for (331), this (302), that (290), python (279), with (260), can (164), data (156), are (151), but (115), from (108), words (105), when (105), more (103), text (103), name (102), your (101), not (100), __getattribute__ (98), which (95), print (92), attribute (92), self (85), code (80), about (80), have (80), first (76), how (75), mark (75), use (74), table (74), there (73), model (73), now (72), test (68), like (66), word (66), one (65), our (65), class (65), all (64), into (62), see (62), each (60), they (60), #pytorch (60), what (59), #tensorflow (56), calling (55), get (53), using (52), time (51), also (51), def (51), other (50), 2026 (50), project (49), has (48), __getattr__ (48), may (47), second (47), argument (47), let (47), than (46), vocabulary (46), training (46), will (45), some (45), where (45), most (44), item (43), object (43), descriptor (43), _data (43), language (42), out (41), these (41), only (39), feature (39), set (39), different (38), them (38), then (38), was (38), return (38), method (38), instance (38), here (37), make (37), button (37), game (37), arrow (37), pointstable (37), need (36), many (36), two (36), people (35), before (35), used (35), value (35), work (35), through (34), their (34), production (34), talk (32), just (32), new (32), any (32), special (32), open (31), look (31), its (31), things (31), __getitem__ (31), news (30), call (30), don (30), pycharm (29), __init__ (29), might (29), whether (29), same (29), very (29), import (28), well (28), natural (28), ag_news_train (28), because (28), matrix (27), community (27), both (27), research (27), languages (26), utc (26), example (26), validation (26), would (26), think (26), calls (26), every (25), even (25), support (25), bow (25), between (25), third (25), non (25), streamlit (25), learn (24), dot (24), title (24), something (24), since (24), methods (24), learning (24), notation (24), path (24), __get__ (24), show (23), way (23), memory (23), count (23), deployment (23), does (22), while (22), thing (22), doesn (22), true (22), output (22), want (22), know (22), block (22), real (21), find (21), bag (21), dataframe (21), take (21), rather (21), models (21), performance (21), ag_news_val (21), across (21), article (21), should (21), write (21), access (21), cursor (21), pycon (20), full (20), help (20), been (20), built (20), column (20), single (20), much (20), english (20), however (20), note (20), stats (20), attributes (20), blog (19), user (19), over (19), dataset (19), uses (19), right (19), label (19), train (19), why (19), checks (19), pep (19), key (19), llm (19), generated (19), issue (18), coding (18), simple (18), next (18), case (18), such (18), category (18), after (18), without (18), frog (18), defined (18), fourth (18), stephen (17), who (17), still (17), based (17), information (17), loss (17), order (17), making (17), list (17), cleaned (17), end (17), within (17), file (17), read (16), tools (16), nlp (16), steps (16), terms (16), apply_string_cleaning (16), string (16), had (16), hand (16), form (16), following (16), start (16), type (16), google (16), buttons (16), bold (16), programming (15), conference (15), directly (15), create (15), try (15), building (15), stop (15), always (15), own (15), features (15), being (15), actually (15), int64 (15), apply (15), back (15), series (15), torch (15), build (15), approach (15), later (15), files (15), choose (15), click (15), age (15), too (15), dynamic (15), sapir (15), whorf (15), council (15), sql (15), fifth (15), descriptors (15), square (15), development (14), line (14), choice (14), lot (14), long (14), post (14), improve (14), 000 (14), values (14), large (14), high (14), issues (14), story (14), hard (14), down (14), number (14), top (14), str (14), super (14), standard (14), really (14), types (14), fetch (14), mobile (14), framework (14), markdown (14), sub (14), works (13), articles (13), users (13), bit (13), april (13), looking (13), fun (13), check (13), saw (13), means (13), working (13), lemmatization (13), already (13), strong (13), vector (13), space (13), per (13), called (13), multiple (13), actual (13), source (13), ag_news_test (13), did (13), add (13), function (13), above (13), techniques (13), going (13), len (13), world (13), specific (13), never (13), strip (13), systems (13), often (13), examples (13), feel (13), api (13), last (13), textual (13), mssql (13), __set__ (13), technology (12), applications (12), important (12), experience (12), similar (12), step (12), patterns (12), signal (12), text_clean (12), size (12), input (12), dictionary (12), names (12), returns (12), via (12), printing (12), common (12), analysis (12), another (12), side (12), writing (12), none (12), those (12), projects (12), come (12), days (12), image (12), release (12), place (12), someone (12), packaging (12), sentinel (12), sprints (12), else (12), edge (12), scale (12), dict (12), datadescriptor (12), nondatadescriptor (12), tuple (12), therefore (12), removal (11), linux (11), past (11), year (11), sign (11), short (11), few (11), classification (11), compared (11), result (11), gives (11), raw (11), description (11), transform (11), less (11), makes (11), process (11), token (11), spacy (11), batch_size (11), batch (11), least (11), install (11), processing (11), keynote (11), quiz (11), app (11), normal (11), enterprise (11), console (11), typer (11), attributeerror (11), raises (11), brackets (11), tech (10), developers (10), group (10), europython (10), part (10), run (10), pipeline (10), inspect (10), started (10), added (10), structure (10), corpus (10), creates (10), though (10), appears (10), accuracy (10), results (10), cleaning (10), modeling (10), pattern (10), four (10), epoch (10), acc (10), val (10), columns (10), appear (10), texts (10), once (10), basic (10), getting (10), copy (10), llms (10), core (10), section (10), times (10), creating (10), soon (10), missing (10), link (10), together (10), talking (10), fallback (10), say (10), server (10), alpha (10), zero (9), three (9), stack (9), complex (9), software (9), libraries (9), join (9), best (9), sure (9), yet (9), couple (9), done (9), github (9), idf (9), representation (9), faster (9), baseline (9), entirely (9), useful (9), meaning (9), idea (9), produce (9), confusion (9), around (9), particular (9), predicted (9), title_clean (9), description_clean (9), pass (9), focused (9), index (9), running (9), forward (9), created (9), convert (9), datasets (9), outputs (9), optimization (9), frequency (9), present (9), cases (9), inverse (9), provides (9), good (9), resources (9), objects (9), pandas (9), face (9), package (9), version (9), during (9), debugging (9), include (9), frogs (9), human (9), humans (9), isn (9), syntax (9), cogs (9), cpython (9), library (9), ecosystem (9), serving (9), verbose (9), testingattributeaccess (9), com (8), tryton (8), speed (8), better (8), michael (8), red (8), david (8), podcast (8), made (8), worth (8), especially (8), comes (8), free (8), preprocessing (8), chart (8), complexity (8), encoding (8), documents (8), topic (8), handles (8), document (8), problems (8), completely (8), distinct (8), to_numpy (8), predictions (8), earlier (8), decisions (8), final (8), science (8), noise (8), except (8), default (8), fit (8), sets (8), again (8), rows (8), forms (8), base (8), int (8), longer (8), ways (8), looks (8), turn (8), giving (8), countvectorizernews (8), contains (8), remove (8), replace (8), regex (8), fix (8), ever (8), anything (8), hugging (8), select (8), management (8), integration (8), certain (8), architectures (8), content (8), despite (8), happens (8), either (8), patch (8), released (8), tell (8), level (8), got (8), revisit (8), behind (8), interface (8), changes (8), gets (8), care (8), deal (8), express (8), turkish (8), native (8), speaker (8), yes (8), incremental (8), team (8), maybe (8), database (8), driver (8), polars (8), defines (8), experimentation (8), graphs (8), frameworks (8), keyerror (8), foundation (7), testing (7), pablo (7), planet (7), volunteer (7), speakers (7), person (7), runs (7), meaningful (7), jupyter (7), vectors (7), easier (7), tasks (7), range (7), off (7), unique (7), amount (7), collection (7), sports (7), business (7), error (7), counts (7), lemmatise_text (7), text_no_stopwords (7), tfidfvectorizernews (7), applied (7), finally (7), including (7), rich (7), could (7), vectorization (7), small (7), categories (7), classes (7), takes (7), intermediate (7), understand (7), keras (7), years (7), machine (7), rarely (7), ones (7), explore (7), change (7), lines (7), canada (7), enter (7), pip (7), sentiment (7), naturally (7), awakening (7), friendly (7), points (7), tricks (7), trick (7), inbox (7), interaction (7), clicked (7), green (7), images (7), documentation (7), force (7), probably (7), garbage (7), collector (7), dialog (7), everyone (7), steering (7), others (7), trends (7), whatever (7), apache (7), pyarrow (7), accepted (7), abi (7), lite (7), mature (7), execution (7), plain (7), courses (7), kate (7), sarah (7), hierarchy (7), scikit (6), chat (6), ide (6), matters (6), low (6), doing (6), video (6), chris (6), great (6), today (6), please (6), keep (6), everything (6), commit (6), difference (6), recent (6), day (6), tutorial (6), numerical (6), seen (6), problem (6), entire (6), individual (6), task (6), dimensionality (6), sometimes (6), context (6), treated (6), mean (6), split (6), generate (6), learned (6), reducing (6), 544 (6), frequently (6), easily (6), larger (6), batches (6), detection (6), field (6), assign (6), functions (6), ag_news_train_cv (6), hidden (6), layer (6), against (6), labels (6), update (6), optimizer (6), control (6), sum (6), hidden_layer_size (6), multiclassclassificationmodel (6), ndarray (6), max (6), originally (6), term (6), reduce (6), noticed (6), particularly (6), straightforward (6), move (6), row (6), reason (6), share (6), apostrophes (6), choosing (6), automatically (6), involves (6), ability (6), easy (6), expect (6), goes (6), increase (6), games (6), bell (6), his (6), contributors (6), sprint (6), sweet (6), delivered (6), actions (6), handle (6), intuitive (6), sending (6), pick (6), haven (6), cake (6), interesting (6), gendered (6), friend (6), agent (6), recently (6), directory (6), enough (6), asking (6), pyconus (6), expressions (6), stuff (6), were (6), increasingly (6), option (6), sentences (6), utf (6), nvarchar (6), execute (6), fetches (6), operations (6), novel (6), custom (6), flexibility (6), infrastructure (6), regular (6), bool (6), multiline (6), define (6), freeze (6), django (5), engineering (5), microsoft (5), tutorials (5), anywhere (5), peter (5), web (5), van (5), pythonic (5), thoughts (5), attendees (5), showing (5), session (5), meet (5), sessions (5), along (5), minutes (5), further (5), exploring (5), yourself (5), virtual (5), setup (5), focus (5), improving (5), statistics (5), approaches (5), cost (5), far (5), allocation (5), depending (5), address (5), capture (5), resulting (5), simply (5), sentence (5), matter (5), understanding (5), generate_predictions (5), baseline_model (5), remove_stopwords (5), toarray (5), away (5), improvements (5), stable (5), shape (5), instead (5), almost (5), countvectorizer (5), recall (5), quite (5), unlikely (5), smaller (5), builds (5), previous (5), consistently (5), tokens (5), additional (5), needs (5), original (5), passes (5), mapping (5), likely (5), numpy (5), gave (5), typically (5), weaknesses (5), pretty (5), itself (5), heavily (5), press (5), characters (5), below (5), reference (5), until (5), implementing (5), tool (5), search (5), environment (5), ready (5), setting (5), exist (5), tooling (5), runtime (5), experimenting (5), messages (5), negative (5), beyond (5), concepts (5), whole (5), army (5), boy (5), paper (5), mario (5), snakes (5), members (5), design (5), track (5), break (5), given (5), play (5), able (5), cloud (5), nintendo (5), translation (5), happen (5), consider (5), allow (5), state (5), associated (5), quickly (5), month (5), put (5), variable (5), saying (5), won (5), mış (5), tense (5), came (5), pronouns (5), known (5), follow (5), general (5), computer (5), local (5), amanda (5), questions (5), technical (5), httpxyz (5), interactions (5), trust (5), slop (5), algorithms (5), companies (5), everybody (5), fetching (5), workloads (5), million (5), devices (5), strengths (5), comprehensive (5), gpu (5), adoption (5), dim (5), total_words (5), flags (5), word_count (5), nprinting (5), deals (5), printout (5), maintainers (5), beta (5), tips (4), weekly (4), tim (4), papers (4), richard (4), ideas (4), paul (4), patrick (4), nick (4), mike (4), matthew (4), little (4), jonathan (4), john (4), jeff (4), cross (4), daniel (4), platform (4), ben (4), haskell (4), shifts (4), roles (4), avoid (4), talks (4), excellent (4), opportunity (4), hours (4), contribution (4), interested (4), topics (4), leaving (4), covered (4), remains (4), latent (4), prediction (4), receives (4), semantics (4), handling (4), alternatives (4), headlines (4), practical (4), sparse (4), related (4), contribute (4), nothing (4), treats (4), require (4), inference (4), main (4), errors (4), genuine (4), must (4), point (4), applying (4), remain (4), separate (4), train_text_classification_model (4), ag_news_train_tfidf (4), get_feature_names_out (4), big (4), vocabulary_ (4), splits (4), joins (4), needed (4), removing (4), requires (4), distinguishing (4), football (4), tensors (4), relu (4), input_size (4), num_classes (4), validation_features (4), validation_labels (4), x_val (4), y_val (4), batch_labels (4), continue (4), selection (4), simplest (4), lower (4), tokenization (4), classic (4), practice (4), generally (4), includes (4), stand (4), fact (4), vocab (4), total (4), false (4), headline (4), guess (4), position (4), possessive (4), supports (4), introduce (4), compact (4), quick (4), become (4), give (4), samples (4), ag_news_all (4), action (4), pipelines (4), trying (4), immediately (4), advantage (4), mistakes (4), effect (4), workflow (4), parameter (4), popular (4), helps (4), efficiently (4), spam (4), wrote (4), binary (4), creation (4), diving (4), represents (4), starting (4), offers (4), her (4), prince (4), gains (4), leads (4), combat (4), area (4), land (4), tolls (4), rom (4), kaeru (4), tame (4), kane (4), naru (4), posts (4), ticket (4), friends (4), mon (4), week (4), display (4), feedback (4), provide (4), script (4), future (4), https (4), dynamically (4), behaviour (4), discussion (4), existing (4), pressure (4), typed (4), isinstance (4), options (4), async (4), referring (4), comments (4), didn (4), him (4), various (4), french (4), living (4), london (4), says (4), psf (4), social (4), wonderful (4), room (4), coming (4), behavior (4), five (4), reading (4), showed (4), azure (4), deep (4), windows (4), conversion (4), noticeably (4), mssql_python (4), conn (4), datetimeoffset (4), hardware (4), reader (4), usage (4), recordbatch (4), industry (4), graph (4), reinforcement (4), vision (4), requirements (4), aren (4), implementations (4), eager (4), head (4), __name__ (4), round (4), panel (4), raise (4), digit (4), exists (4), chars_no_space (4), blocks (4), rules (4), __setattr__ (4), traceback (4), decide (4), subscriptable (4), bracket (4), sqlalchemy (4), dedicated (4), 772 (4), elected (4), percent (4), request (3), łukasz (3), langa (3), dev (3), william (3), ram (3), taylor (3), simon (3), weblog (3), rodrigo (3), robert (3), beginners (3), mitchell (3), hugo (3), kemenade (3), reliable (3), society (3), roy (3), greenfeld (3), dan (3), christian (3), carl (3), bruno (3), awesome (3), feed (3), powered (3), event (3), ask (3), slots (3), clicking (3), involved (3), helping (3), knows (3), organized (3), view (3), pro (3), scientific (3), friction (3), fundamentals (3), weighting (3), performs (3), discover (3), goal (3), representations (3), captures (3), dependencies (3), trade (3), computational (3), close (3), limitations (3), established (3), fraction (3), majority (3), matrices (3), cheap (3), regardless (3), limit (3), fundamental (3), sequence (3), describe (3), overfitting (3), easiest (3), classify (3), misclassified (3), boundary (3), reflect (3), exactly (3), locked (3), consistent (3), seems (3), extra (3), deploy (3), ag_news_val_tfidf (3), semantically (3), failure (3), official (3), uninformative (3), frequent (3), generalize (3), limiting (3), pipe (3), filtered_texts (3), punctuation (3), correctly (3), identify (3), detail (3), commonly (3), shows (3), company (3), ag_news_val_cv (3), specify (3), trained (3), picture (3), alone (3), backward (3), weights (3), evaluation (3), mode (3), loop (3), starts (3), indexed (3), converted (3), dataloader (3), neural (3), dimensional (3), allows (3), highest (3), becomes (3), fc2 (3), fc1 (3), optim (3), module (3), num_epochs (3), floattensor (3), train_loader (3), criterion (3), train_loss (3), correct_train (3), total_train (3), metrics (3), val_outputs (3), match (3), predicted_labels (3), successful (3), reusable (3), possible (3), relevant (3), reduces (3), preserving (3), major (3), vectorizer (3), introducing (3), carry (3), unseen (3), appeared (3), depends (3), everywhere (3), equivalent (3), stemming (3), slower (3), extremely (3), nltk (3), clean (3), displayed (3), axis (3), corresponding (3), wrap (3), counting (3), updated (3), notebook (3), spaces (3), wide (3), openai (3), apis (3), lets (3), variables (3), contractions (3), window (3), mostly (3), preparing (3), summary (3), perform (3), train_test_split (3), to_pandas (3), 120 (3), loading (3), versions (3), interpreter (3), inline (3), unexpected (3), specialized (3), broad (3), emails (3), presence (3), positive (3), neutral (3), efficient (3), effective (3), represented (3), implement (3), thousands (3), recording (3), fits (3), fall (3), wondered (3), delarin (3), somewhat (3), curses (3), towards (3), asymmetry (3), favorite (3), managed (3), seeing (3), upgrade (3), battle (3), nantendo (3), visit (3), door (3), intelligent (3), whom (3), japanese (3), releases (3), narrative (3), progress (3), galindo (3), salgado (3), stage (3), july (3), live (3), spy (3), 2025 (3), submit (3), submitted (3), hello (3), inputs (3), pressed (3), displaying (3), statement (3), welcome (3), apps (3), response (3), rest (3), system (3), costs (3), requests (3), links (3), readable (3), genuinely (3), unnatural (3), conversation (3), aware (3), scope (3), course (3), explicit (3), sync (3), due (3), some_func (3), love (3), notice (3), kind (3), subconsciously (3), having (3), reveal (3), sex (3), obvious (3), temporary (3), limits (3), clear (3), speak (3), factory (3), implementation (3), marimo (3), pair (3), dialogs (3), thought (3), sentinels (3), breaking (3), passed (3), introduced (3), some_arg (3), member (3), advice (3), booth (3), effort (3), student (3), moved (3), seattle (3), grateful (3), mentor (3), conversations (3), themselves (3), fork (3), lots (3), pygrunn (3), engagement (3), engage (3), influence (3), agents (3), assumption (3), happening (3), plenty (3), domains (3), elsewhere (3), reach (3), engaging (3), wanted (3), found (3), spikes (3), looked (3), internal (3), scenarios (3), felix (3), odbc (3), tested (3), fast (3), current (3), connect (3), fetchall (3), json (3), serialization (3), temporal (3), significantly (3), tables (3), available (3), arrow_reader (3), lazy (3), duckdb (3), recordbatchreader (3), 8192 (3), peak (3), throughput (3), arrow_batch (3), programs (3), proven (3), optimized (3), onnx (3), torchserve (3), decision (3), override (3), mlops (3), deploying (3), cutting (3), tpu (3), loops (3), tensor (3), verdict (3), pre (3), curve (3), definition (3), min (3), 200 (3), _count_single_file (3), prefixed (3), style (3), reading_time_min (3), chars (3), box (3), cyan (3), html (3), markers (3), inner (3), readme (3), weird (3), forum (3), fails (3), initially (3), protocol (3), __dict__ (3), finds (3), stands (3), keys (3), numbers (3), sqlite (3), stuck (3), seats (3), 803 (3), body (3), authority (3), hypothesis (2), wes (2), mason (2), wayne (2), vinay (2), sajip (2), babu (2), twisted (2), labs (2), thomas (2), guest (2), occurrence (2), digital (2), sumana (2), stefan (2), simeon (2), seth (2), larson (2), sebastian (2), girão (2), serrão (2), robin (2), wilson (2), docs (2), péter (2), collaborative (2), guis (2), board (2), celery (2), müller (2), kennedy (2), nicolas (2), neil (2), schemenauer (2), driscoll (2), marc (2), luke (2), plant (2), technologies (2), das (2), kumar (2), kelly (2), kay (2), julien (2), cook (2), joe (2), jeremy (2), rocha (2), gustavo (2), guido (2), rossum (2), grant (2), graham (2), françois (2), flavio (2), duncan (2), math (2), art (2), davy (2), dave (2), school (2), command (2), corey (2), christoph (2), calvin (2), consulting (2), automating (2), artem (2), armin (2), ronacher (2), antonio (2), cuni (2), anton (2), andy (2), alexandre (2), alex (2), michiel (2), perl (2), planets (2), rss (2), huge (2), success (2), connecting (2), worst (2), necessary (2), interact (2), ensure (2), smoothly (2), role (2), volunteers (2), manage (2), website (2), seasoned (2), perfect (2), notebooks (2), environments (2), minimal (2), converts (2), meaningfully (2), efficiency (2), viewer (2), assistant (2), reaching (2), generating (2), substantially (2), representing (2), continuous (2), embeddings (2), several (2), considering (2), sized (2), vast (2), sparsity (2), consume (2), stored (2), harder (2), phenomenon (2), curse (2), ignores (2), limitation (2), unordered (2), discarding (2), dog (2), man (2), identical (2), events (2), question (2), answering (2), serious (2), tells (2), properties (2), test_predictions (2), ag_news_test_tfidf (2), test_accuracy (2), crucially (2), kept (2), risk (2), choices (2), estimate (2), optimistic (2), hardest (2), encouraging (2), overall (2), contributing (2), manageable (2), 9475 (2), 9243 (2), 5000 (2), improved (2), considerably (2), weighted (2), importance (2), relatively (2), fed (2), pension (2), government (2), tfidfvectorizer (2), max_features (2), 20000 (2), fixed (2), scoring (2), discards (2), tail (2), rare (2), arbitrary (2), switching (2), conveniently (2), lemmatized (2), filters (2), remaining (2), loaded (2), is_stop (2), doc (2), grammatical (2), parser (2), processes (2), preserve (2), download (2), en_core_web_sm (2), load (2), discussed (2), ratio (2), packages (2), sense (2), contexts (2), equal (2), epochs (2), complete (2), held (2), prints (2), breakdown (2), confused (2), informative (2), rmsprop (2), switches (2), monitor (2), expects (2), network (2), transformations (2), whose (2), architecture (2), walk (2), utils (2), tensordataset (2), linear (2), train_features (2), train_labels (2), learning_rate (2), train_labels_indexed (2), validation_labels_indexed (2), arrays (2), x_train (2), y_train (2), longtensor (2), train_dataset (2), parameters (2), batch_features (2), eval (2), no_grad (2), val_loss (2), val_predicted (2), correct_val (2), total_val (2), train_acc (2), val_acc (2), written (2), asked (2), jetbrains (2), encoded (2), applies (2), completion (2), statistical (2), target (2), suited (2), widely (2), reduction (2), component (2), flexible (2), tend (2), fitted (2), preserves (2), unknown (2), perhaps (2), onto (2), phrases (2), grams (2), adjacent (2), resolve (2), rule (2), truncation (2), linguistic (2), produces (2), kinds (2), adding (2), solid (2), quality (2), appropriate (2), advanced (2), hover (2), left (2), occur (2), distribution (2), typical (2), frequencies (2), visualization (2), corner (2), visualizing (2), sort (2), settings (2), drop (2), moving (2), expected (2), array (2), 108 (2), defaults (2), payments (2), stores (2), ensures (2), ran (2), meaningless (2), sklearn (2), removed (2), barrick (2), gold (2), acquires (2), nine (2), cent (2), stake (2), celtic (2), canadian (2), racism (2), chief (2), insert (2), cell (2), letter (2), compatible (2), prompt (2), inside (2), possessives (2), strips (2), patterns_to_remove (2), whitespace (2), normalization (2), noted (2), displays (2), detailed (2), dataframes (2), prevent (2), inspection (2), increases (2), gradually (2), finished (2), datasetdict (2), num_rows (2), containing (2), load_dataset (2), online (2), pypi (2), dependency (2), supported (2), components (2), navigation (2), spend (2), effects (2), keeps (2), experiments (2), lists (2), visibility (2), sizes (2), hints (2), potential (2), relies (2), recognize (2), determine (2), sorted (2), simplicity (2), underlying (2), scales (2), collections (2), significant (2), advantages (2), consequence (2), elements (2), records (2), element (2), checking (2), simpler (2), overwhelmed (2), occurs (2), near (2), arrangement (2), tracking (2), expressed (2), unpack (2), relative (2), modern (2), early (2), under (2), thanks (2), keeping (2), light (2), accomplish (2), defeat (2), princess (2), background (2), minute (2), croakian (2), sablé (2), mandola (2), transforming (2), eventually (2), chests (2), actively (2), kingdoms (2), factions (2), health (2), difficult (2), discovered (2), stat (2), reward (2), rumors (2), scrolling (2), perspective (2), folks (2), dug (2), assembly (2), fairly (2), engine (2), challenge (2), inc (2), wario (2), assist (2), references (2), adventure (2), products (2), needing (2), developed (2), chapter (2), pigs (2), hemingway (2), definitely (2), spoilers (2), played (2), cartridge (2), screen (2), playing (2), recommend (2), stumbled (2), iván (2), youtube (2), researching (2), zelda (2), popping (2), life (2), knowledge (2), published (2), wait (2), crowd (2), sales (2), planned (2), hey (2), stages (2), episode (2), watch (2), slow (2), text_input (2), number_input (2), min_value (2), max_value (2), allowing (2), triggers (2), immediate (2), structured (2), cover (2), executed (2), switch (2), img_url_1 (2), placehold (2), 150 (2), img_url_2 (2), caption (2), expanded (2), providing (2)

Text of the page (random words):
s at zero this signal to noise issue is something we ll return to advantages of the bag of words model the bag of words model has remained a staple in nlp for good reason its greatest strength is its simplicity because text is represented as a collection of word counts the approach is easy to understand and straightforward to implement making it a natural baseline before reaching for more complex architectures beyond simplicity bow is computationally efficient as you saw above the underlying math is lightweight which means it scales well to large text collections without demanding significant computing resources for tasks where the presence of specific words is sufficient to capture meaning with sentiment analysis and topic categorization being the clearest examples it remains a highly effective tool applications of bag of words like many nlp approaches the bag of words model can be applied to many natural language problems these potential applications include document classification where encoded documents are sorted into predefined categories a classic example of this is automatically sorting incoming news articles into distinct categories such as sports politics or technology as we ll see in the project we do in this post sentiment analysis where the presence of certain words strongly indicates the overall tone of a text allows models to easily determine whether a piece of writing expresses a positive negative or neutral sentiment if you re interested in learning more about bow and other approaches to sentiment analysis you can see a prior blog post i wrote on this topic spam detection which relies heavily on bow to identify and filter out unwanted emails or messages by learning to recognize the distinct high frequency word patterns characteristic of spam retrieval systems where it helps to efficiently find the most relevant documents from an immense corpus based on a user s search query topic modeling which aims to group similar text vectors in order to discover and extract the hidden latent topics present within a large collection of documents as you can see the number of potential applications is broad making bag of words modeling a popular first approach to natural language problems why use pycharm for nlp pycharm is particularly well suited to bag of words modeling because it supports the iterative detail oriented workflow that text processing requires as you ll soon see building a reliable bow pipeline involves multiple steps such as cleaning text tokenizing vectorizing and validating outputs and pycharm s code intelligence makes each of these smoother autocompletion parameter hints and quick navigation through specialized nlp libraries reduce friction when experimenting with different vectorizer settings and help you understand how each component behaves debugging and data inspection are equally important here since small preprocessing mistakes can have an outsized effect on results pycharm lets you step through your code and examine intermediate states of things such as token lists and vocabulary at runtime making it straightforward to verify that your feature extraction is working as intended this visibility is especially useful when diagnosing issues like unexpected vocabulary sizes or missing terms pycharm also supports exploratory work through its excellent jupyter notebook integration and scientific tooling bow modeling often involves trying different preprocessing strategies and observing their effects immediately so the ability to run code interactively and inspect outputs inline is a genuine advantage combined with built in virtual environment and package management support this keeps experiments reproducible and well organized as projects grow pycharm s refactoring tools project navigation and version control integration help manage the added complexity bow models rarely exist in isolation and they re often embedded in larger ml pipelines in such contexts pycharm s features for working with larger applications mean you spend less time managing code and more time improving your models setting up the project to see these components in action let s build an actual bag of words project we ll use a classic text classification dataset and the ag news dataset and then use the model to classify news articles into one of four categories world sports business or science technology to get started in pycharm open the projects and files tool window and select new new project since this is a data science project we can use pycharm s built in jupyter project type which sets up a sensible default structure for us during project configuration you ll be asked to choose a python interpreter by default pycharm uses uv and lets you select from a range of python versions though all major dependency management systems are supported pip anaconda pipenv poetry and hatch every project is automatically created with an attached virtual environment so your setup will be ready to go each time you reopen the project with the project configured we can install our dependencies via the python packages tool window simply search for a package by name select it from the list and install your desired version directly into the virtual environment you can also see the same information about the package you d find on pypi directly within the ide for this project we ll need pandas and numpy along with datasets from hugging face scikit learn pytorch and spacy implementing bag of words with pycharm there are many versions of this dataset online we ll be using one of the versions hosted on hugging face hub loading and preparing the data we ll use hugging face s datasets package to download this dataset from datasets import load_dataset ag_news_all load_dataset sh0416 ag_news this gives us a hugging face datasetdict object if we look at it we can see it contains a training dataset with 120 000 news articles and a test dataset containing 7 600 articles ag_news_all datasetdict train dataset features label title description num_rows 120000 test dataset features label title description num_rows 7600 as we ll be training a model we also need a validation set we ll convert the training and test sets to pandas dataframes and use the train_test_split method from scikit learn to create the validation set from the training data import pandas as pd from sklearn model_selection import train_test_split ag_news_train ag_news_all train to_pandas ag_news_test ag_news_all test to_pandas ag_news_train ag_news_val train_test_split ag_news_train test_size 0 1 random_state 456 stratify ag_news_train label print f training set len ag_news_train samples print f validation set len ag_news_val samples we now have a validation set with 12 000 articles and a training set with 108 000 articles training set 108000 samples validation set 12000 samples for those of you new to machine learning you might be wondering why we need all of these different datasets the reason for this is to make sure we have a good idea that our model will generalize well and perform as expected on unseen data the training set is the only data the model ever learns from directly the validation set is used to monitor how the model is performing on unseen data as we make modeling decisions such as choosing how many epochs to train for how large to make the hidden layer or which preprocessing steps to apply we ll see all of this later this means that we look at validation performance repeatedly while building the model and this increases the risk that our choices gradually become tuned to the quirks of that particular split this is why we need a third set the test set which we keep completely locked away until we ve finished all modeling decisions and want a single unbiased estimate of how well our model will perform on new data using the test set for anything other than this final evaluation would give us an overly optimistic picture of our model s real world performance let s now inspect our datasets pycharm pro has a lot of built in features that make working with dataframes easier a few of which we ll see soon in this dataframe we have three columns the article title and description the article text and the label indicating which of the four news categories the article belongs to you can open any of the dataframe cells in the value editor to see its full text or widen the column to prevent truncation both of which are useful for a quick visual inspection at the top of each column pycharm displays column statistics giving you an at a glance summary of the data switching from compact to detailed mode via show column statistics gives you rich summary statistics about each column and saves you from writing a lot of pandas boilerplate to get it from these statistics we can see that our training set is evenly split across the news categories which is very handy when training a model we can also see that some headlines and descriptions are not unique which may introduce noise when classifying these duplicates the first step in preparing the data is basic string cleaning which normalizes the text and reduces meaningless token variation for instance without cleaning natural and natural would be treated as two separate vocabulary entries as we noted earlier we ll apply four cleaning steps lowercasing punctuation removal number removal and whitespace normalization there are different string cleaning steps you can apply depending on the language and use case but for english language texts these tend to be very standard let s go ahead and write a function to do this def apply_string_cleaning dataset pd series pd series patterns_to_remove r a za z s cleaned dataset str lower for pattern in patterns_to_remove cleaned cleaned str replace pattern regex true cleaned cleaned str replace r s regex true str strip return cleaned ag_news_train title_clean apply_string_cleaning ag_news_train title ag_news_train description_clean apply_string_cleaning ag_news_train description this mostly works but there s one issue the regex strips apostrophes entirely turning contractions like you re into you re and possessives like canada s into canada s the cleanest fix is a regex that preserves apostrophes in contractions while removing possessive endings but this is not the most enjoyable thing to write by hand this is where pycharm s built in ai assistant comes in open the chat window via the ai chat icon on the right hand side of the ide and enter the following prompt can you please alter the apply_string_cleaning function so that it retains apostrophes inside words when they re used for contractions e g you re but removes them when they re used for possessives e g canada s into canada the notation lets you reference specific files or objects in your ide without copying and pasting code into the prompt including jupyter variables like datasets and functions i ran this against claude sonnet 4 5 though jetbrains ai supports a wide range of models from openai anthropic google and xai as well as open models via ollama lm studio and openai compatible apis below is the updated function it returned def apply_string_cleaning dataset pd series pd series cleaned dataset str lower remove possessive apostrophes word s word this pattern matches letter s s word boundary cleaned cleaned str replace r w s b r 1 regex true remove all non letter characters except apostrophes within words cleaned cleaned str replace r a za z s regex true clean up any apostrophes at the start or end of words cleaned cleaned str replace r s s regex true remove multiple spaces and trim cleaned cleaned str replace r s regex true str strip return cleaned ag_news_train title_clean apply_string_cleaning ag_news_train title ag_news_train description_clean apply_string_cleaning ag_news_train description we can insert this into our jupyter notebook directly by clicking on insert snippet as jupyter cell in the ai chat once we run this updated function on our raw text we get the correct result text text_clean don t stand for racism football chief don t stand for racism football chief canada s barrick gold acquires nine per cent stake in celtic resources canadian press canada barrick gold acquires nine per cent stake in celtic resources canadian press we can see the contraction don t is correctly preserved in the first example but the possessive canada s has been removed we apply this to both the training and validation datasets using the same function so that the cleaning is consistent across both splits ag_news_val title_clean apply_string_cleaning ag_news_val title ag_news_val description_clean apply_string_cleaning ag_news_val description creating the bag of words model now that we have clean text we need to build our vocabulary and encode it we ll use scikit learn s countvectorizer for this from sklearn feature_extraction text import countvectorizer countvectorizernews countvectorizer countvectorizernews fit ag_news_train text_clean ag_news_train_cv countvectorizernews transform ag_news_train text_clean toarray the process has two distinct steps first fit scans the training data and builds a vocabulary by identifying every unique word and assigning it a fixed index position for example government column 8 901 the result is a mapping of 59 544 unique words which you can think of as the column headers for our eventual matrix second transform uses that vocabulary to convert each headline into a numerical vector counting how many times each vocabulary word appears and placing that count at the corresponding index position the reason these are two separate steps is important when we later process our validation and test data we ll call transform using the vocabulary learned from the training set this ensures that all three splits share a consistent feature space if we re ran fit on the test data we d get a different vocabulary and the model s predictions would be meaningless with the vectorizer fitted and our training data transformed we can start exploring what we ve actually built let s first take a look at the vocabulary countvectorizer stores it as a dictionary mapping each word to its index position accessible via vocabulary_ countvectorizernews vocabulary_ fed 18461 up 55833 with 58324 pension 38929 defaults 13156 citing 9475 failure 18077 of 36704 two 54804 big 5269 airlines 1139 to 53531 make 31397 payments 38686 their 52947 len countvectorizernews vocabulary_ 59544 this confirms that our vocabulary contains 59 544 unique words browsing through it you can start to guess what kinds of terms appear frequently in the different types of news country names feature heavily in the world news category terms like football and cricket in the sports news category terms like profit and losses in the business news category and company names like google and microsoft in the science technology category next let s inspect the feature matrix itself ag_news_train_cv is a numpy array with one row per headline and one column per vocabulary word giving us a matrix of shape 108 000 59...