Text Generation
Transformers
Safetensors
Arabic
English
llama
translation
egyptian-arabic
english
arabic
small-language-model
slm
tiny-lm
chatml
scaling-study
text-generation-inference
Instructions to use oddadmix/Emhotob-5M-Egyptian-English-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use oddadmix/Emhotob-5M-Egyptian-English-v1 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="oddadmix/Emhotob-5M-Egyptian-English-v1")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("oddadmix/Emhotob-5M-Egyptian-English-v1") model = AutoModelForCausalLM.from_pretrained("oddadmix/Emhotob-5M-Egyptian-English-v1", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use oddadmix/Emhotob-5M-Egyptian-English-v1 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "oddadmix/Emhotob-5M-Egyptian-English-v1" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "oddadmix/Emhotob-5M-Egyptian-English-v1", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/oddadmix/Emhotob-5M-Egyptian-English-v1
- SGLang
How to use oddadmix/Emhotob-5M-Egyptian-English-v1 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "oddadmix/Emhotob-5M-Egyptian-English-v1" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "oddadmix/Emhotob-5M-Egyptian-English-v1", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "oddadmix/Emhotob-5M-Egyptian-English-v1" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "oddadmix/Emhotob-5M-Egyptian-English-v1", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use oddadmix/Emhotob-5M-Egyptian-English-v1 with Docker Model Runner:
docker model run hf.co/oddadmix/Emhotob-5M-Egyptian-English-v1
Egyptian-English 5M SFT + card + eval (scaling study)
Browse files- README.md +129 -0
- config.json +32 -0
- eval_bidirectional.json +223 -0
- generation_config.json +10 -0
- model.safetensors +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +15 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,129 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- ar
|
| 5 |
+
- en
|
| 6 |
+
base_model: oddadmix/Emhotob-5M
|
| 7 |
+
pipeline_tag: text-generation
|
| 8 |
+
library_name: transformers
|
| 9 |
+
tags:
|
| 10 |
+
- translation
|
| 11 |
+
- egyptian-arabic
|
| 12 |
+
- english
|
| 13 |
+
- arabic
|
| 14 |
+
- small-language-model
|
| 15 |
+
- slm
|
| 16 |
+
- tiny-lm
|
| 17 |
+
- chatml
|
| 18 |
+
- scaling-study
|
| 19 |
+
metrics:
|
| 20 |
+
- bleu
|
| 21 |
+
- chrf
|
| 22 |
+
---
|
| 23 |
+
|
| 24 |
+
# Emhotob-5M-Egyptian-English-v1 โ Bidirectional English โ Egyptian Arabic (~5M params)
|
| 25 |
+
|
| 26 |
+
A **5.08M-parameter** model that translates **both ways** between **English** and
|
| 27 |
+
**Egyptian colloquial Arabic (ุงูู
ุตุฑูุฉ ุงูุนุงู
ูุฉ)**. A single set of weights serves both
|
| 28 |
+
directions; a direction-specific system prompt selects which way to translate.
|
| 29 |
+
|
| 30 |
+
Finetuned from [`oddadmix/Emhotob-5M`](https://huggingface.co/oddadmix/Emhotob-5M),
|
| 31 |
+
a tiny Llama-architecture base (hidden size 128, 5 layers, 4 heads, tied embeddings).
|
| 32 |
+
|
| 33 |
+
> **Scaling study.** This runs the exact recipe of
|
| 34 |
+
> [`oddadmix/50M-Egyptian-English-v1`](https://huggingface.co/oddadmix/50M-Egyptian-English-v1)
|
| 35 |
+
> on a base ~10ร smaller. At 5M the model produces **fluent-looking target-language text and
|
| 36 |
+
> switches language/register correctly, but the content is often wrong** โ it drifts and
|
| 37 |
+
> repeats heavily. Cross-lingual EnglishโEgyptian is measurably harder at this size than the
|
| 38 |
+
> intra-Arabic MSAโEgyptian pair. It is a scaling demonstration, not a usable translator; use
|
| 39 |
+
> the 50M sibling for real translation.
|
| 40 |
+
|
| 41 |
+
## Evaluation
|
| 42 |
+
|
| 43 |
+
Evaluated on a **deterministic held-out set of 3,000 pairs** (`seed=42`), decoded
|
| 44 |
+
**greedily** (`do_sample=False`, no repetition penalty), scored with sacreBLEU:
|
| 45 |
+
|
| 46 |
+
| Direction | sacreBLEU | chrF |
|
| 47 |
+
|---|---:|---:|
|
| 48 |
+
| **English โ Egyptian** | **0.83** | 13.15 |
|
| 49 |
+
| **Egyptian โ English** | **1.87** | 14.84 |
|
| 50 |
+
|
| 51 |
+
The saved weights are the best checkpoint by validation loss (`eval_loss = 3.367`, epoch 3 of 3).
|
| 52 |
+
For reference, the 50M sibling scores BLEU ~24 (enโegy) / ~34 (egyโen) with the same data and eval.
|
| 53 |
+
|
| 54 |
+
### Example translations
|
| 55 |
+
|
| 56 |
+
Real greedy-decoded outputs from the held-out set (rough โ this is a 5M model):
|
| 57 |
+
|
| 58 |
+
**English โ Egyptian**
|
| 59 |
+
|
| 60 |
+
| English input | Model output (Egyptian) | Reference |
|
| 61 |
+
|---|---|---|
|
| 62 |
+
| So what should I get you? Should I get you this or that? | ูุนูู ุฅููุ ุฅูู ุฏูุ ุฅูู ุฏูุ โฆ ููููู ุนูู ุญุงุฌุฉ | ุทุจ ุฃุฌูุจ ูู ุฅููุ ุฃุฌูุจ ูู ุฏู ููุง ุฏูุ |
|
| 63 |
+
|
| 64 |
+
**Egyptian โ English**
|
| 65 |
+
|
| 66 |
+
| Egyptian input | Model output (English) | Reference |
|
| 67 |
+
|---|---|---|
|
| 68 |
+
| ุทุจ ุฃุฌูุจ ูู ุฅููุ ุฃุฌูุจ ูู ุฏู ููุง ุฏูุ | What's this? What's this, what's this? | So what should I get you? Should I get you this or that? |
|
| 69 |
+
|
| 70 |
+
The model emits well-formed Egyptian / English surface text and picks the right language for the
|
| 71 |
+
direction, but frequently loses the source meaning. 20 samples per direction with references are
|
| 72 |
+
in [`eval_bidirectional.json`](./eval_bidirectional.json).
|
| 73 |
+
|
| 74 |
+
## Usage
|
| 75 |
+
|
| 76 |
+
ChatML format. **Pick the system prompt for the direction you want:**
|
| 77 |
+
|
| 78 |
+
```python
|
| 79 |
+
import torch
|
| 80 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 81 |
+
|
| 82 |
+
model_id = "oddadmix/Emhotob-5M-Egyptian-English-v1"
|
| 83 |
+
tok = AutoTokenizer.from_pretrained(model_id)
|
| 84 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, dtype=torch.bfloat16).to("cuda").eval()
|
| 85 |
+
|
| 86 |
+
SYS_TO_EGY = "ุฃูุช ู
ุชุฑุฌู
ู
ุญุชุฑู. ุชุฑุฌู
ุงููุต ุงูุฅูุฌููุฒู ุฅูู ุงูููุฌุฉ ุงูู
ุตุฑูุฉ ุงูุนุงู
ูุฉ."
|
| 87 |
+
SYS_TO_EN = "You are a professional translator. Translate the Egyptian Arabic text into English."
|
| 88 |
+
|
| 89 |
+
def translate(text: str, system: str) -> str:
|
| 90 |
+
prompt = (
|
| 91 |
+
f"<|im_start|>system\n{system}<|im_end|>\n"
|
| 92 |
+
f"<|im_start|>user\n{text.strip()}<|im_end|>\n"
|
| 93 |
+
f"<|im_start|>assistant\n"
|
| 94 |
+
)
|
| 95 |
+
ids = tok(prompt, return_tensors="pt", add_special_tokens=False).to(model.device)
|
| 96 |
+
if tok.bos_token_id is not None: # training prepends BOS
|
| 97 |
+
bos = torch.tensor([[tok.bos_token_id]], device=model.device)
|
| 98 |
+
ids["input_ids"] = torch.cat([bos, ids["input_ids"]], dim=1)
|
| 99 |
+
ids["attention_mask"] = torch.cat([torch.ones_like(bos), ids["attention_mask"]], dim=1)
|
| 100 |
+
out = model.generate(**ids, max_new_tokens=256, do_sample=False,
|
| 101 |
+
eos_token_id=tok.eos_token_id, pad_token_id=tok.pad_token_id)
|
| 102 |
+
return tok.decode(out[0, ids["input_ids"].size(1):], skip_special_tokens=True).strip()
|
| 103 |
+
|
| 104 |
+
print(translate("I will be fine, don't worry", SYS_TO_EGY))
|
| 105 |
+
```
|
| 106 |
+
|
| 107 |
+
## Training
|
| 108 |
+
|
| 109 |
+
- **Base model:** `oddadmix/Emhotob-5M` (Llama arch, hidden 128, 5 layers, 4 heads, vocab 32000,
|
| 110 |
+
tied embeddings; **5,080,704 params** after resizing for 2 ChatML tokens)
|
| 111 |
+
- **Dataset:** `oddadmix/egyptian-translation-dataset-2.9-openai-batch` (135,282 rows,
|
| 112 |
+
`English`/`Arabic` columns; `Arabic` = Egyptian colloquial)
|
| 113 |
+
- **Method:** HuggingFace `Trainer`, ChatML, **prompt-masked cross-entropy** (loss only on the
|
| 114 |
+
assistant turn). Each row is exploded into **two** training examples (one per direction).
|
| 115 |
+
Two ChatML special tokens (`<|im_start|>`, `<|im_end|>`) were added and embeddings resized.
|
| 116 |
+
- **Hyperparameters:** 3 epochs ยท effective batch 64 ยท LR 3e-4 (cosine, 5% warmup) ยท
|
| 117 |
+
bf16 ยท max length 1024 ยท `load_best_model_at_end` on `eval_loss`.
|
| 118 |
+
- **Split:** 132,282 train / 3,000 deterministic held-out (`seed=42`), scored both directions.
|
| 119 |
+
|
| 120 |
+
## Limitations
|
| 121 |
+
|
| 122 |
+
- A 5M model near the **emergence threshold**: fluent surface text but frequent meaning loss,
|
| 123 |
+
drift, and repetition โ especially on longer or rarer inputs.
|
| 124 |
+
- Gender is disambiguated only from context; ambiguous inputs may default one way.
|
| 125 |
+
- For fluent translation use `oddadmix/50M-Egyptian-English-v1`.
|
| 126 |
+
|
| 127 |
+
## License
|
| 128 |
+
|
| 129 |
+
Apache-2.0, inherited from the base model.
|
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlamaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"dtype": "bfloat16",
|
| 9 |
+
"eos_token_id": 32001,
|
| 10 |
+
"head_dim": 32,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 128,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 384,
|
| 15 |
+
"max_position_embeddings": 2048,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 4,
|
| 19 |
+
"num_hidden_layers": 5,
|
| 20 |
+
"num_key_value_heads": 2,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-06,
|
| 24 |
+
"rope_parameters": {
|
| 25 |
+
"rope_theta": 10000,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"tie_word_embeddings": true,
|
| 29 |
+
"transformers_version": "5.12.1",
|
| 30 |
+
"use_cache": false,
|
| 31 |
+
"vocab_size": 32002
|
| 32 |
+
}
|
eval_bidirectional.json
ADDED
|
@@ -0,0 +1,223 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eval_size": 3000,
|
| 3 |
+
"seed": 42,
|
| 4 |
+
"generation": {
|
| 5 |
+
"do_sample": false,
|
| 6 |
+
"repetition_penalty": 1.0,
|
| 7 |
+
"no_repeat_ngram_size": 0
|
| 8 |
+
},
|
| 9 |
+
"directions": {
|
| 10 |
+
"en2egy": {
|
| 11 |
+
"bleu": 0.8284798113158242,
|
| 12 |
+
"chrf": 13.151559692353024,
|
| 13 |
+
"samples": [
|
| 14 |
+
{
|
| 15 |
+
"source": "So what should I get you? Should I get you this or that? Just imagine the scene: sitting now, wearing a sleeveless undershirt, eating the best piece of meat you'll ever have in your life.",
|
| 16 |
+
"reference": "ุทุจ ุฃุฌูุจ ูู ุฅููุ ุฃุฌูุจ ูู ุฏู ููุง ุฏูุ ุชุฎูู ูุฏู ุงูู
ูุธุฑุ ูุงุนุฏ ุฏูููุชู ูุงุจุณ ูุงููุฉ ุจุญู
ุงูุงุช ูุจุชุงูู ุฃุญูู ุญุชุฉ ูุญู
ุฉ ู
ู
ูู ุชูููุง ูู ุญูุงุชู.",
|
| 17 |
+
"hypothesis": "ูุนูู ุฅููุ ุฅูู ุฏูุ ุฅูู ุฏูุ ุฃูุง ู
ุด ููููุนุ ููููู ุนูู ุญุงุฌุฉุ ููููู ุนูู ุญุงุฌุฉุ ุฃูุง ู
ุด ููููุน."
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"source": "Prayer niches (mihrabs)",
|
| 21 |
+
"reference": "ู
ุญุงุฑูุจ ุงูุตูุงุฉ",
|
| 22 |
+
"hypothesis": "ู
ู
ูู ุชูุฌู"
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"source": "Cardamom, grains of paradise, cardamom, black pepper, and a little salt. We'll just sprinkle these on, with some bay leaves, and we'll add a big lemon so it doesn't darken โ here's the big lemon.",
|
| 26 |
+
"reference": "ุญุจูุงู ููููู ุงุณูุฏ ูุดููู ู
ูุญุ ููุฑุดูู
ุนูููุง ู
ุน ูุฑู ููุฑูุ ูููุญุทููุง ููู
ููุฉ ูุจูุฑุฉ ุนุดุงู ููููุง ู
ุง ูุบู
ูุด - ุงูู ููู
ููุฉ ูุจูุฑุฉ.",
|
| 27 |
+
"hypothesis": "ูุง ุฌู
ุงุนุฉุ ุฃูููุ ูู ุณูุฉ ูุงูุชูุง ู
ุด ููููุนูุงุ ูุจุนุฏูู ูุงูููุงุ ูุจุนุฏูู ูุงูููุงุ ูุจุนุฏูู ูุงูููุง. ุนุงูุฒูุง ูุฑูุญูุงุ ุนุดุงู ูุฏู ูุชุจูู ุฒู ู
ุง ูู."
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"source": "Man, these things happen. Sometimes Liverpool loses the match, and other times Real Madrid wins it.",
|
| 31 |
+
"reference": "ูุง ุนู
ุงูุญุงุฌุงุช ุฏู ุจุชุญุตูุ ุณุงุนุงุช ูููุฑุจูู ุจุชุฎุณุฑ ุงูู
ุงุชุด ูุณุงุนุงุช ุฑูุงู ู
ุฏุฑูุฏ ุจููุณุจ.",
|
| 32 |
+
"hypothesis": "ูุง ุฌู
ุงุนุฉุ ุงููุงุณ ุจุชููู ุฅู ูู ู
ุงุชุด ู
ุงูุดุณุชุฑ ุณูุชูุ ูุฏู ูุงู ุจูุนู
ููุง ูุฏู."
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"source": "I will be fine, don't worry",
|
| 36 |
+
"reference": "ุงูุง ูุจูู ูููุณุฉุ ู
ุง ุชููููุด",
|
| 37 |
+
"hypothesis": "ุฃูุง ู
ุด ุนุงุฑูุ ุฃูุง ู
ุด ููููุน"
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"source": "I want grilled chicken, kofta on charcoal, and Pepsi.",
|
| 41 |
+
"reference": "ุงูุง ุนุงูุฒ ูุฑุงุฎ ู
ุดููุฉ ูููุชุฉ ุนูู ุงููุญู
ูุจูุจุณู.",
|
| 42 |
+
"hypothesis": "ุนุงูุฒ ุฃูููู ุนูู ุญุงุฌุฉุ ูุจุนุฏูู ูุฑูุญ."
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"source": "Your age",
|
| 46 |
+
"reference": "ุณูู",
|
| 47 |
+
"hypothesis": "ู
ู
ูู"
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"source": "By the way, today, it's interesting to see how these two teams will handle the result, each playing in their own way. Whether it's Klopp trying to give hope, even if it's just 10%, that we can achieve the impossible in Madrid against a very tough opponent, or, if not, weโll at least have a clear, explicit, and realistic goal so the season doesn't slip away from us. A win would make all Liverpool fans completely quiet with Jurgen Klopp and give them an injection of hope until the end of the season, at least. Unlike United, who desperately need an automatic, clear, and decisive reaction starting from the Europa League against Real Betis. These tough, historic results stay in our memories regardless of your position in the leagueโwe all remember them.",
|
| 51 |
+
"reference": "ุนูู ููุฑูุ ุงูููุงุฑุฏู ุงูู
ูุถูุน ููุจูู interesting ูุดูู ุงููุฑูููู ุฏูู ููุชุนุงู
ููุง ู
ุน ุงููุชูุฌู ุงุฒุงูุ ููู ูุงุญุฏ ูููุนุจ ุจุทุฑููุชู. ุณูุงุก ูููุจ ููุญุงูู ูุฏู ุฃู
ูุ ุญุชู ูู ุจูุณุจุฉ ูกู ูชุ ุฅููุง ููุฏุฑ ูุนู
ู ุงูู
ุณุชุญูู ูู ู
ุฏุฑูุฏ ูุฏุงู
ุฎุตู
ุชููู ุฌุฏุงูุ ุฃู ูู ูุฃุ ูุจูู ุนูู ุงูุฃูู ุนูุฏูุง ูุฏู ูุงุถุญ ูุตุฑูุญ ููุงูุนู ุนุดุงู ุงูู
ูุณู
ู
ุง ูุถูุนุด ู
ููุง. ุงูููุฒ ููุฎูู ุฌู
ููุฑ ูููุฑุจูู ูุณูุช ุชู
ุงู
ุงู ู
ุน ููุฑุฌู ูููุจ ููุฏููู
ุฌุฑุนุฉ ุฃู
ู ูุญุฏ ููุงูุฉ ุงูู
ูุณู
ุนูู ุงูุฃูู. ุนูู ุนูุณ ูููุงูุชุฏ ุงููู ู
ุญุชุงุฌ ุฑุฏ ูุนู ุณุฑูุน ููุงุถุญ ูุตุฑูุญ ุงุจุชุฏุงุกู ู
ู ุงูููุฑูุจุง ููุฌ ูุฏุงู
ุฑูุงู ุจูุชูุณ. ุงููุชุงุฆุฌ ุงูุตุนุจุฉ ูุงูุชุงุฑูุฎูุฉ ุฏู ุจุชูุถู ูู ุงูุฐุงูุฑุฉ ู
ูู
ุง ูุงู ู
ุฑูุฒู ูู ุงูุฏูุฑู โ ูููุง ูุงูุฑูููุง.",
|
| 52 |
+
"hypothesis": "ุนูู ููุฑุฉุ ูู ู
ุด ุนุงุฑู ูู ูุงู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ูู ูุงู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ูู ูุงู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ูุงู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุฃู ูู ุงูุฏูุฑูุ ุฃู ูู ุงูุฏูุฑูุ ุฃู ูู ๏ฟฝ๏ฟฝูุฏูุฑูุ ุฃู ูู ุงูุฏูุฑูุ ุฃู ูู ุงูุฏูุฑูุ ุฃู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู ูู ุงูุฏูุฑูุ ุฃู ูู ูุงู"
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"source": "So, do we have to pay here?",
|
| 56 |
+
"reference": "ูุนูู ุฅุญูุง ูุงุฒู
ูุฏูุน ููุงุ",
|
| 57 |
+
"hypothesis": "ูุนูู ุนุงูุฒ ุฃูููู ุนูู ููุงุ"
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"source": "I think we will know that.",
|
| 61 |
+
"reference": "ุฃุธู ุฅุญูุง ููุนุฑู ุฏู.",
|
| 62 |
+
"hypothesis": "ุฃูุง ู
ุด ุนุงุฑู ุฅู ุฏู ู
ุด ููุญุตู."
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"source": "Option please, excuse me, thank you.",
|
| 66 |
+
"reference": "ุฃูุจุดู ูู ุณู
ุญุชุ ุจุนุฏ ุฅุฐููุ ุดูุฑุงู.",
|
| 67 |
+
"hypothesis": "ูุง ุฌู
ุงุนุฉุ ุฃูุง ู
ุด ููููู."
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"source": "That's what people say, I don't know.",
|
| 71 |
+
"reference": "ุฏู ุงููู ุงููุงุณ ุจุชููููุ ู
ุนุฑูุด.",
|
| 72 |
+
"hypothesis": "ุฏู ู
ุด ููุจูู ุจูููููุง ุฅูู."
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"source": "No sir, he's going to clean up this mess. He might come if France fails in this tournament. No sir, they supported him until he asked 'which cup is this?' I mean, this is Mbappe, he'll play in 2026. On paper, yes, but if he gets eliminated early, like in the round of 16, God knows. He was eliminated from Euro 2021, true, but when they gave him a chance to continue, he did really well.",
|
| 76 |
+
"reference": "ูุง ูุง ุจุงุดุงุ ูู ูููุถู ุงููููุฉ ุฏู. ู
ู
ูู ููุฌู ูู ูุฑูุณุง ูุดูุช ูู ุงูุจุทููุฉ ุฏู. ูุง ูุง ุจุงุดุงุ ูุงููุง ู
ุนุงู ูุญุฏ ู
ุง ูุงู ูู ุงููุงุณ ุฏู ุฅููุ ูุนูู ุฏู ู
ุจุงุจูุ ูููุนุจ ูู 2026. ุนูู ุงููุฑู ุขูุ ุจุณ ูู ุฎุฑุฌ ุจุฏุฑู ุฒู ุฏูุฑ ุงูุณุชุงุดุฑุ ุงููู ุฃุนูู
. ูู ุฎุฑุฌ ู
ู ุงูููุฑู 2021ุ ุตุญุ ุจุณ ูู
ุง ุงุฏููู ูุฑุตุฉ ููู
ูุ ุฃุซุจุช ููุณู ุฌุงู
ุฏ.",
|
| 77 |
+
"hypothesis": "ูุง ูุง ูุงุ ูู ูุงู ุจูููู ุฅู ูู ุงูุขุฎุฑ ููุจูู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู. ูู ูุงู ุจูููู ุฅู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุจุณ ูู ูุงู ุจูููู ุฅู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุจุณ ูู ูุงู ุจูููู ุฅู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุจุณ ูู ูุงู ุจูููู ุฅู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุจุณ ูู ูุงู ุจูููู ุฅู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุจุณ ูู ูุงู ุจูููู ุฅู ูู ู
ุงุชุด ูู ุงูุฏูุฑูุ ุจุณ ูู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ุจูููู ุฅู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู ูุงู ูู ุงูุฏูุฑู ุงูุฅูุฌููุฒู."
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"source": "Let me tell you, I ask them to parse what's underlined, and I say, 'the champion raises the cup.' 'Champion' is the subject. Then, Malak (a student) says to me, 'the subject is marked with a kasra (wrong grammatical case)'โcan you imagine? I swear that's what happened!",
|
| 81 |
+
"reference": "ุจุต ุจููุ ุงูุง ุจุณุฃููู
ูุนุฑุจูุง ุงููู ุชุญุชูุง ุฎุทุ ูุจููู ููู
ุงูุจุทู ุจูุฑูุน ุงููุงุณุ ูุนูู 'ุงูุจุทู' ูู ุงูู
ุจุชุฏุฃ. ุชููู
ู
ููุ ุชุฎูููุ ุชููู ูู: 'ุงูู
ุจุชุฏุฃ ู
ุฑููุน ุจุงููุณุฑ'! ุชุฎููู! ูุงููู ุญุตู ูุฏู.",
|
| 82 |
+
"hypothesis": "ุจุตุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉ. ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉ. ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉ."
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"source": "What's my standard anyway? Forget about his standards, the awards' standards, and all that stuff.",
|
| 86 |
+
"reference": "ูู ุงูู ุงุตูุง ุงูู
ุนูุงุฑ ุจุชุงุนูุ ุณูุจู ุจูู ู
ู ู
ุนุงููุฑู ูู
ุนุงููุฑ ุงูุฌูุงูุฒ ููู ุงูููุงู
ุฏู.",
|
| 87 |
+
"hypothesis": "ุฅูู ุงููู ุจูุญุตูุ ุฃูุง ู
ุด ููููุนุ ุจุณ ูู ุญุฏ ุจูุนู
ูู."
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"source": "I wish I could go back to being young, to the days when I didn't carry worries like I do now.",
|
| 91 |
+
"reference": "ูุง ุฑูุช ุฃุฑุฌุน ุตุบูุฑ ุชุงูู ูุฃูุงู
ู
ุงููุชุด ุดุงูู ูู
ุฒู ุฏูููุชู.",
|
| 92 |
+
"hypothesis": "ูุง ุฑุจุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉ."
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"source": "My question is, what kind of work is Guardiola doing with this one and that one? Where did he succeed? In Barcelona, he definitely succeeded with both. You can only succeed in Germany? No, I mean, can you appreciate what this man has done? This man worked hard and did something amazing.",
|
| 96 |
+
"reference": "ุฃูุง ุณุคุงูู ูู ุฅูู ุงูุดุบู ุงููู ุฌูุงุฑุฏูููุง ุจูุนู
ูู ู
ุน ุฏู ูู
ุน ุฏูุ ูู ูุฌุญ ูููุ ูู ุจุฑุดูููุฉ ูุฌุญ ูุนูุงู ู
ุน ุงูุงุชููู. ูุนูู ุงููุฌุงุญ ุจุณ ูู ุฃูู
ุงููุงุ ูุงุ ุฃูุง ูุตุฏู ุชูุฏูุฑูุง ุงูุฑุงุฌู ุฏู ุนู
ู ุงููุ ุงูุฑุงุฌู ุฏู ุชุนุจ ูุนู
ู ุญุงุฌู ูุจูุฑุฉ.",
|
| 97 |
+
"hypothesis": "ุจุต ูุง ุฌู
ุงุนุฉุ ุฅูู ุฏูุ ุฅูู ุฏูุ ูู ุฅููุ ูู ุฅููุ ูู ุฅููุ ูู ุฅููุ ูู ุฅูู ูู ุฅููุ ูู ุฅูู ูู ุฅููุ ูู ุฅูู ูู ุฅููุ ูู ุฅูู ูู ุฅููุ"
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"source": "Check yourself, Mr. Harry. You destroyed Tottenham Clubโlooks like you were the problem. They might win this year now that Kane is gone or whatever.",
|
| 101 |
+
"reference": "ุฑุงุฌุน ููุณู ูุง ุนู
ูุงุฑูุ ุงูุช ุฎุฑุจุช ูุงุฏู ุชูุชููุงู
ุ ุดููู ููุช ุงูุช ุงูุณุจุจ. ูู
ูู ููุณุจูุง ุงูุณูุฉ ุฏู ุจุนุฏ ู
ุง ููู ู
ุดู ููุง ุฃู ุญุงุฌุฉ.",
|
| 102 |
+
"hypothesis": "ููุง ูุง ุฌู
ุงุนุฉุ ุฃูุง ู
ุด ููููู ุนูู ุญุงุฌุฉ. ุฃูุง ู
ุด ููููุนุ ุฏู ู
ุด ููุจูู ูู ูู ุญุงุฌุฉ."
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"source": "Let's go, I'll get up, go to Rouh El Nagham, and call you on the phone.",
|
| 106 |
+
"reference": "ููุงุ ูููู
ุฃุฑูุญ ุฑูุญ ุงููุบู
ููููู
ู ูู ุงูุชููููู.",
|
| 107 |
+
"hypothesis": "ููุง ูุง ุจุงุดุงุ ุฃูุง ู
ุด ูููููุ ุฃูุง ู
ุด ููููู."
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"source": "And he'll make the kids deal with a mean stepmother.",
|
| 111 |
+
"reference": "ูููุฎูู ุงูุนูุงู ูุชุนุงู
ููุง ู
ุน ู
ุฑุงุช ุฃุจ ุดุฑูุฑุฉ.",
|
| 112 |
+
"hypothesis": "ูุจุนุฏูู ูู ุงููู ุจูุญุตู ูู ูู ุญุชุฉ."
|
| 113 |
+
}
|
| 114 |
+
]
|
| 115 |
+
},
|
| 116 |
+
"egy2en": {
|
| 117 |
+
"bleu": 1.8709159904583916,
|
| 118 |
+
"chrf": 14.84143953330391,
|
| 119 |
+
"samples": [
|
| 120 |
+
{
|
| 121 |
+
"source": "ุทุจ ุฃุฌูุจ ูู ุฅููุ ุฃุฌูุจ ูู ุฏู ููุง ุฏูุ ุชุฎูู ูุฏู ุงูู
ูุธุฑุ ูุงุนุฏ ุฏูููุชู ูุงุจุณ ูุงููุฉ ุจุญู
ุงูุงุช ูุจุชุงูู ุฃุญูู ุญุชุฉ ูุญู
ุฉ ู
ู
ูู ุชูููุง ูู ุญูุงุชู.",
|
| 122 |
+
"reference": "So what should I get you? Should I get you this or that? Just imagine the scene: sitting now, wearing a sleeveless undershirt, eating the best piece of meat you'll ever have in your life.",
|
| 123 |
+
"hypothesis": "What's this? What's this, what's this? What's this, what's this, what's this? I'm telling you, what's this, and I'm telling you, what's this?"
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"source": "ู
ุญุงุฑูุจ ุงูุตูุงุฉ",
|
| 127 |
+
"reference": "Prayer niches (mihrabs)",
|
| 128 |
+
"hypothesis": "And the situation is something"
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"source": "ุญุจูุงู ููููู ุงุณูุฏ ูุดููู ู
ูุญุ ููุฑุดูู
ุนูููุง ู
ุน ูุฑู ููุฑูุ ูููุญุทููุง ููู
ููุฉ ูุจูุฑุฉ ุนุดุงู ููููุง ู
ุง ูุบู
ูุด - ุงูู ููู
ููุฉ ูุจูุฑุฉ.",
|
| 132 |
+
"reference": "Cardamom, grains of paradise, cardamom, black pepper, and a little salt. We'll just sprinkle these on, with some bay leaves, and we'll add a big lemon so it doesn't darken โ here's the big lemon.",
|
| 133 |
+
"hypothesis": "They're saying, and we'll take a bit, and you'll take a bit, and you'll take a bit, and you'll take a bit of a bit, and you'll take a bit of a bit."
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"source": "ูุง ุนู
ุงูุญุงุฌุงุช ุฏู ุจุชุญุตูุ ุณุงุนุงุช ูููุฑุจูู ุจุชุฎุณุฑ ุงูู
ุงุชุด ูุณุงุนุงุช ุฑูุงู ู
ุฏุฑูุฏ ุจููุณุจ.",
|
| 137 |
+
"reference": "Man, these things happen. Sometimes Liverpool loses the match, and other times Real Madrid wins it.",
|
| 138 |
+
"hypothesis": "Hey, this is the one who wouldn't take the match, and they're the one who wouldn't be a bit."
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"source": "ุงู๏ฟฝ๏ฟฝ ูุจูู ูููุณุฉุ ู
ุง ุชููููุด",
|
| 142 |
+
"reference": "I will be fine, don't worry",
|
| 143 |
+
"hypothesis": "I'm not sure, I'll tell you, I'll tell you, it's not a bit."
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"source": "ุงูุง ุนุงูุฒ ูุฑุงุฎ ู
ุดููุฉ ูููุชุฉ ุนูู ุงููุญู
ูุจูุจุณู.",
|
| 147 |
+
"reference": "I want grilled chicken, kofta on charcoal, and Pepsi.",
|
| 148 |
+
"hypothesis": "I'm not sure you're not saying it's a bit."
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"source": "ุณูู",
|
| 152 |
+
"reference": "Your age",
|
| 153 |
+
"hypothesis": "He's something"
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"source": "ุนูู ููุฑูุ ุงูููุงุฑุฏู ุงูู
ูุถูุน ููุจูู interesting ูุดูู ุงููุฑูููู ุฏูู ููุชุนุงู
ููุง ู
ุน ุงููุชูุฌู ุงุฒุงูุ ููู ูุงุญุฏ ูููุนุจ ุจุทุฑููุชู. ุณูุงุก ูููุจ ููุญุงูู ูุฏู ุฃู
ูุ ุญุชู ูู ุจูุณุจุฉ ูกู ูชุ ุฅููุง ููุฏุฑ ูุนู
ู ุงูู
ุณุชุญูู ูู ู
ุฏุฑูุฏ ูุฏุงู
ุฎุตู
ุชููู ุฌุฏุงูุ ุฃู ูู ูุฃุ ูุจูู ุนูู ุงูุฃูู ุนูุฏูุง ูุฏู ูุงุถุญ ูุตุฑูุญ ููุงูุนู ุนุดุงู ุงูู
ูุณู
ู
ุง ูุถูุนุด ู
ููุง. ุงูููุฒ ููุฎูู ุฌู
ููุฑ ูููุฑุจูู ูุณูุช ุชู
ุงู
ุงู ู
ุน ููุฑุฌู ูููุจ ููุฏููู
ุฌุฑุนุฉ ุฃู
ู ูุญุฏ ููุงูุฉ ุงูู
ูุณู
ุนูู ุงูุฃูู. ุนูู ุนูุณ ูููุงูุชุฏ ุงููู ู
ุญุชุงุฌ ุฑุฏ ูุนู ุณุฑูุน ููุงุถุญ ูุตุฑูุญ ุงุจุชุฏุงุกู ู
ู ุงูููุฑูุจุง ููุฌ ูุฏุงู
ุฑูุงู ุจูุชูุณ. ุงููุชุงุฆุฌ ุงูุตุนุจุฉ ูุงูุชุงุฑูุฎูุฉ ุฏู ุจุชูุถู ูู ุงูุฐุงูุฑุฉ ู
ูู
ุง ูุงู ู
ุฑูุฒู ูู ุงูุฏูุฑู โ ูููุง ูุงูุฑูููุง.",
|
| 157 |
+
"reference": "By the way, today, it's interesting to see how these two teams will handle the result, each playing in their own way. Whether it's Klopp trying to give hope, even if it's just 10%, that we can achieve the impossible in Madrid against a very tough opponent, or, if not, weโll at least have a clear, explicit, and realistic goal so the season doesn't slip away from us. A win would make all Liverpool fans completely quiet with Jurgen Klopp and give them an injection of hope until the end of the season, at least. Unlike United, who desperately need an automatic, clear, and decisive reaction starting from the Europa League against Real Betis. These tough, historic results stay in our memories regardless of your position in the leagueโwe all remember them.",
|
| 158 |
+
"hypothesis": "The same thing is that they're not saying, 'We'll be a bit, and I'm not saying that they're saying, 'I'm not saying, 'I'm not saying, 'I'm not saying,' and I'm not saying, 'I'm not saying, 'I'm not saying,' and I'm not saying that they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying they're saying the"
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"source": "ูุนูู ุฅุญูุง ูุงุฒู
ูุฏูุน ููุงุ",
|
| 162 |
+
"reference": "So, do we have to pay here?",
|
| 163 |
+
"hypothesis": "So, what's this?"
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"source": "ุฃุธู ุฅุญูุง ููุนุฑู ุฏู.",
|
| 167 |
+
"reference": "I think we will know that.",
|
| 168 |
+
"hypothesis": "I want to tell you, we're not saying."
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"source": "ุฃูุจุดู ูู ุณู
ุญุชุ ุจุนุฏ ุฅุฐููุ ุดูุฑุงู.",
|
| 172 |
+
"reference": "Option please, excuse me, thank you.",
|
| 173 |
+
"hypothesis": "And if you're not saying, I'm not saying, I'm not saying."
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"source": "ุฏู ุงููู ุงููุงุณ ุจุชููููุ ู
ุนุฑูุด.",
|
| 177 |
+
"reference": "That's what people say, I don't know.",
|
| 178 |
+
"hypothesis": "This is the one who wouldn't tell you."
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"source": "ูุง ูุง ุจุงุดุงุ ูู ูููุถู ุงููููุฉ ุฏู. ู
ู
ูู ููุฌู ูู ูุฑูุณุง ูุดูุช ูู ุงูุจุทููุฉ ุฏู. ูุง ูุง ุจุงุดุงุ ูุงููุง ู
ุนุงู ูุญุฏ ู
ุง ูุงู ูู ุงููุงุณ ุฏู ุฅููุ ูุนูู ุฏู ู
ุจุงุจูุ ูููุนุจ ูู 2026. ุนูู ุงููุฑู ุขูุ ุจุณ ูู ุฎุฑุฌ ุจุฏุฑู ุฒู ุฏูุฑ ุงูุณุชุงุดุฑุ ุงููู ุฃุนูู
. ูู ุฎุฑุฌ ู
ู ุงูููุฑู 2021ุ ุตุญุ ุจุณ ูู
ุง ุงุฏููู ูุฑุตุฉ ููู
ูุ ุฃุซุจุช ููุณู ุฌุงู
ุฏ.",
|
| 182 |
+
"reference": "No sir, he's going to clean up this mess. He might come if France fails in this tournament. No sir, they supported him until he asked 'which cup is this?' I mean, this is Mbappe, he'll play in 2026. On paper, yes, but if he gets eliminated early, like in the round of 16, God knows. He was eliminated from Euro 2021, true, but when they gave him a chance to continue, he did really well.",
|
| 183 |
+
"hypothesis": "No, they're not saying this is the one who would be a bit. What's this, what's this, what's this? If you want to tell you, it's not a bit, but it's not a bit. If you're not saying, 'I'm not saying, 'I'm not saying, 'I'm telling you, 'I'm telling you,' and I'm not saying this isn't a bit.'"
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"source": "ุจุต ุจููุ ุงูุง ุจุณุฃููู
ูุนุฑุจูุง ุงููู ุชุญุชูุง ุฎุทุ ูุจููู ููู
ุงูุจุทู ุจูุฑูุน ุงููุงุณุ ูุนูู 'ุงูุจุทู' ูู ุงูู
ุจุชุฏุฃ. ุชููู
ู
ููุ ุชุฎูููุ ุชููู ูู: 'ุงูู
ุจุชุฏุฃ ู
ุฑููุน ุจุงููุณุฑ'! ุชุฎููู! ูุงููู ุญุตู ูุฏู.",
|
| 187 |
+
"reference": "Let me tell you, I ask them to parse what's underlined, and I say, 'the champion raises the cup.' 'Champion' is the subject. Then, Malak (a student) says to me, 'the subject is marked with a kasra (wrong grammatical case)'โcan you imagine? I swear that's what happened!",
|
| 188 |
+
"hypothesis": "Look, I'm telling you, I'll take them, and I'll tell you, it's not the same. I'm telling you, it's not something, and I'll take them to the match. So, I'm telling you, it's not a bit of a bit. So, what's they want to tell you, they're saying, 'I'll tell you, 'I'll tell you, 'I'll take it.'"
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"source": "ูู ุงูู ุงุตูุง ุงูู
ุนูุงุฑ ุจุชุงุนูุ ุณูุจู ุจูู ู
ู ู
ุนุงููุฑู ูู
ุนุงููุฑ ุงูุฌูุงูุฒ ููู ุงูููุงู
ุฏู.",
|
| 192 |
+
"reference": "What's my standard anyway? Forget about his standards, the awards' standards, and all that stuff.",
|
| 193 |
+
"hypothesis": "He's not the field? The match is the one who would be a bit, and they're not saying that's the field."
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"source": "ูุง ุฑูุช ุฃุฑุฌุน ุตุบูุฑ ุชุงูู ูุฃูุงู
ู
ุงููุชุด ุดุงูู ูู
ุฒู ุฏูููุชู.",
|
| 197 |
+
"reference": "I wish I could go back to being young, to the days when I didn't carry worries like I do now.",
|
| 198 |
+
"hypothesis": "Hey, I'm not saying that they're saying that they're saying it's a bit."
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"source": "ุฃูุง ุณุคุงูู ูู ุฅูู ุงูุดุบู ุงููู ุฌูุงุฑุฏูููุง ุจูุนู
ูู ู
ุน ุฏู ูู
ุน ุฏูุ ูู ูุฌุญ ูููุ ูู ุจุฑุดูููุฉ ูุฌุญ ูุนูุงู ู
ุน ุงูุงุชููู. ูุนูู ุงููุฌุงุญ ุจุณ ูู ุฃูู
ุงููุงุ ูุงุ ุฃูุง ูุตุฏู ุชูุฏูุฑูุง ุงูุฑุงุฌู ุฏู ุนู
ู ุงููุ ุงูุฑุงุฌู ุฏู ุชุนุจ ูุนู
ู ุญุงุฌู ูุจูุฑุฉ.",
|
| 202 |
+
"reference": "My question is, what kind of work is Guardiola doing with this one and that one? Where did he succeed? In Barcelona, he definitely succeeded with both. You can only succeed in Germany? No, I mean, can you appreciate what this man has done? This man worked hard and did something amazing.",
|
| 203 |
+
"hypothesis": "I'm not saying this is the one who would be a bit, and what's they want to tell you? What's this? What's this, I'm saying this is the one who would be a bit. What's this? What's this, I'm not saying this is the first time, okay? I'm not saying it's a bit, okay?"
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"source": "ุฑุงุฌุน ููุณู ูุง ุนู
ูุงุฑูุ ุงูุช ุฎุฑุจุช ูุงุฏู ุชูุชููุงู
ุ ุดููู ููุช ุงูุช ุงูุณุจุจ. ูู
ูู ููุณุจูุง ุงูุณูุฉ ุฏู ุจุนุฏ ู
ุง ููู ู
ุดู ููุง ุฃู ุญุงุฌุฉ.",
|
| 207 |
+
"reference": "Check yourself, Mr. Harry. You destroyed Tottenham Clubโlooks like you were the problem. They might win this year now that Kane is gone or whatever.",
|
| 208 |
+
"hypothesis": "I want to tell you, you're the one who would be a bit, you're not saying it's a bit. They're not saying that you're not saying it's a bit."
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"source": "ููุงุ ูููู
ุฃุฑูุญ ุฑูุญ ุงููุบู
ููููู
ู ูู ุงูุชููููู.",
|
| 212 |
+
"reference": "Let's go, I'll get up, go to Rouh El Nagham, and call you on the phone.",
|
| 213 |
+
"hypothesis": "Come on, I'll take the same too, and I'll take the moment and see them."
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"source": "ูููุฎูู ุงูุนูุงู ูุชุนุงู
ููุง ู
ุน ู
ุฑุงุช ุฃุจ ุดุฑูุฑุฉ.",
|
| 217 |
+
"reference": "And he'll make the kids deal with a mean stepmother.",
|
| 218 |
+
"hypothesis": "What's they want to tell you, they're saying they're saying."
|
| 219 |
+
}
|
| 220 |
+
]
|
| 221 |
+
}
|
| 222 |
+
}
|
| 223 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 32001,
|
| 5 |
+
"output_attentions": false,
|
| 6 |
+
"output_hidden_states": false,
|
| 7 |
+
"pad_token_id": 1,
|
| 8 |
+
"transformers_version": "5.12.1",
|
| 9 |
+
"use_cache": true
|
| 10 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b7bdc412de37396c3be2d290532a05f888acfcf070c654c83b4737d40dfd742
|
| 3 |
+
size 10166472
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"backend": "tokenizers",
|
| 3 |
+
"bos_token": "<s>",
|
| 4 |
+
"eos_token": "<|im_end|>",
|
| 5 |
+
"extra_special_tokens": [
|
| 6 |
+
"<|im_start|>",
|
| 7 |
+
"<|im_end|>"
|
| 8 |
+
],
|
| 9 |
+
"is_local": false,
|
| 10 |
+
"local_files_only": false,
|
| 11 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 12 |
+
"pad_token": "<pad>",
|
| 13 |
+
"tokenizer_class": "TokenizersBackend",
|
| 14 |
+
"unk_token": "<unk>"
|
| 15 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fda5a230ef60e1fb8a3297ccb58fa6451c4511a187b2eb2858f82782c3b021fd
|
| 3 |
+
size 4792
|