refactor: eagle config rewrite; update doc

[h-guo18]

What does this PR do?

Type of change: documentation

Overview:

• Updated doc;
• Moved config overwriting and d2t loading from main.py -> modify() to simplify the use of mtsp.convert();
• Update README.md accordingly.

Usage

No change

Testing

• Tested eagle confgs are identical before and after refactor:

Before refactor:

eagle config: PretrainedConfig {
  "attention_bias": false,
  "attention_dropout": 0.0,
  "draft_vocab_size": 128256,
  "dtype": "bfloat16",
  "eagle_aux_hidden_state_layer_ids": [
    1,
    12,
    7
  ],
  "has_lm_head": false,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "num_attention_heads": 32,
  "num_hidden_layers": 1,
  "num_key_value_heads": 8,
  "parallel_draft_heads_num_layers": 1,
  "parallel_draft_step": 1,
  "position_embedding_type": "rope",
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 32.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "transformers_version": "4.57.1",
  "use_aux_hidden_state": true,
  "use_input_layernorm_in_first_layer": true,
  "use_last_layernorm": true,
  "use_mtp_layernorm": false,
  "vocab_size": 128256
}

After refactor:

eagle config: PretrainedConfig {
  "attention_bias": false,
  "attention_dropout": 0.0,
  "draft_vocab_size": 128256,
  "dtype": "bfloat16",
  "eagle_aux_hidden_state_layer_ids": [
    1,
    12,
    7
  ],
  "has_lm_head": false,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "num_attention_heads": 32,
  "num_hidden_layers": 1,
  "num_key_value_heads": 8,
  "parallel_draft_heads_num_layers": 1,
  "parallel_draft_step": 1,
  "position_embedding_type": "rope",
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 32.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "transformers_version": "4.57.1",
  "use_aux_hidden_state": true,
  "use_input_layernorm_in_first_layer": true,
  "use_last_layernorm": true,
  "use_mtp_layernorm": false,
  "vocab_size": 128256
}

Before your PR is "Ready for review"

• Make sure you read and follow Contributor guidelines and your commits are signed.
• Is this change backward compatible?: Yes/No
• Did you write any new necessary tests?: Yes/No
• Did you add or update any necessary documentation?: Yes/No
• Did you update Changelog?: Yes/No

Additional Information

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 74.64%. Comparing base (d0b0c0f) to head (69115e7).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #624   +/-   ##
=======================================
  Coverage   74.64%   74.64%           
=======================================
  Files         183      183           
  Lines       18542    18542           
=======================================
  Hits        13840    13840           
  Misses       4702     4702           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.